Night-vision-technology.pdf

AN INTRODUCTION TO

NIGHT VISION TECHNOLOGY

AN INTRODUCTION TO

NIGHT VISION TECHNOLOGY R Hradaynath

DEFENCE RESEARCH & DEVELOPMENT ORGANISATION MINISTRY OF DEFENCE NEW DELHI – 110 011

2002

DRDO Monographs/Special Publications Series An Introduction to Night Vision Technology R Hradaynath

Series Editors Editor-in-Chief

Editors

Mohinder Singh

Ashok Kumar A Saravanan

Asst Editor

Editorial Asst

Ramesh Chander

AK Sen

Production Printing

Cover Design

Marketing

JV Ramakrishna

Vinod Kumari Sharma

RK Dua

SK Tyagi

RK Bhatnagar

© 2002, Defence Scientific Information & Documentation Centre (DESIDOC), Defence R&D Organisation, Delhi-110 054. All rights reserved. Except as permitted under the Indian Copyright Act 1957, no part of this publication may be reproduced, distributed or transmitted, stored in a database or a retrieval system, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the Publisher. The views expressed in the book are those of the author only. The Editors or Publisher do not assume responsibility for the statements/ opinions expressed by the author.

ISBN: 81–86514–10–4 Printed and published by Director, DESIDOC, Metcalfe House, Delhi-110 054.

CONTENTS

Foreword

ix

Preface

xi

Acknowledgements

xv

CHAPTER 1 VISION & HUMAN EYE 1.1

Introduction

1

1.2

Optical parameters of human eye

2

1.3

Information processing by visual system

6

1.4

Overall mechanisms

9

1.4.1

Light stimulus

1.4.2

Threshold Vs intensity functions & contrast

10

1.4.3

Colour

11

1.5

Implications for night vision

12

1

9

CHAPTER 2 SEARCH & ACQUISITION 2.1

Search

15

2.2

Aquisition

16

2.3

Blackwell's approach

18

2.4

Johnson criteria

19

2.5

Display signal-to-noise ratio

20

2.6

Detection with target movement

22

2.7

Probabilities of aquisition

23

2.8

Contrast & acquisition

23

15

CHAPTER 3 THE ENVIRONMENT 3.1

Introduction

29

3.2

Atmospheric absorption & scattering

31

3.2.1

Scattering due to rain & snow

34

3.2.2

Haze & fog

35

3.2.3

Visibility & contrast

35

3.3

Atmosphere modelling

38

3.4

Instruments, night vision & atmospherics

39

29

(vi)

CHAPTER 4 NIGHT ILLUMINATION, REFLECTIVITIES & BACKGROUND 4.1

Night illumination

43

4.1.1

Moonlight

44

4.1.2

Starlight

48

4.2

Reflectivity at night

48

4.3

The background

50

4.4

Effect on design of vision devices

52

43

CHAPTER 5 OPTICAL CONSIDERATIONS 5.1

Introduction

55

5.2

Basic requirements

58

5.2.1

System parameters

59

5.2.2

Design approach

64

5.2.3

Design evaluation

66

5.3

Optical considerations

69

55

CHAPTER 6 PHOTOEMISSION 6.1

Introduction

77

6.2

Photoemission & its theoretical considerations

77

6.2.1

Theoretical considerations

78

6.2.2

Types of photocathodes & their efficiencies

79

6.3

Development of photocathodes

81

6.3.1

Composite photocathodes

81

6.3.2

Alloy photocathodes

81

6.3.3

Alkali photocathodes

82

6.3.4

Negative affinity photocathodes

83

6.3.5

Transferred electron(field assisted) photocathodes

85

6.4

Photocathode response time

87

6.5

Photocathode sensitivity

87

6.6

Dark current in photocathodes

90

6.7

Summary

91

77

CHAPTER 7 PHOSPHORS 7.1

Introduction

93 93

(vii)

7.2

Phosphors

93

7.3

Luminous transitions in a phosphor

94

7.4

Phosphor mechanisms

96

7.5

Reduction of luminescence efficiency

99

7.6

Luminescence decay

99

7.7

Phosphor applications

100

7.8

Phosphor screens

101

7.9

Screen fabrication

103

7.10

Phosphor ageing

104

CHAPTER 8 IMAGE INTENSIFIER TUBES

105

8.1

Introduction

8.2

Fibre optics in image intensifiers

105 108

8.2.1

Concepts of fibre-optics

109

8.2.2

Fibre-optics faceplates

110

8.2.3

Micro-channel plates

114

8.2.4

Fibre-optic image inverters/twisters

117

8.3

Electron optics

117

8.4

General considerations for image intensifier designs

120

8.5

Image intensifier tube types

125

8.5.1

Generation-0 image converter tubes

125

8.5.2

Generation-1 image intensifier tubes

126

8.5.3

Generation-2 image intensifier tubes

127

8.5.4

Generation-2 wafer tube

129

8.5.6

Generation-3 image intensifier tubes

131

8.5.7

Hybrid tubes

132

8.6

Performance of image intensifier tubes

134

8.6.1

Signal-to-noise ratio

134

8.6.2

Consideration of modulation transfer function (MTF)

136

8.6.3

Luminous gain and E.B.I

137

8.6.4

Other parameters

138

8.6.5

A note on production of image intensifier tubes

138

CHAPTER 9 NIGHT VISION INSTRUMENTATION 9.1

Introduction

9.2

Range equation

143 143 144

(viii)

9.3

Experimental lab testing for range evaluation

150

9.4

Field testing

154

9.5

Instrument types

155

Index

163

FOREWORD The author has been one of the main architects in introducing night vision technology to India. He was intimately involved, at a crucial time, in the R&D on this subject leading to the development of a variety of instruments for use by the Armed Forces and their subsequent bulk production through an integrated scheme of technology transfer. The present monograph is a welcome and unique addition to the already existing literature on night vision technology. Besides introducing all the parameters and technologies that comprise this subject, it would also assist a reader to correct his design effort to result in an effective instrument. The development of the subject begins with an understanding of the human eye and vision as also the principles underlying search and acquisition. This study enables one to realise the limits to which human observation is restricted in practice. The study further extends to the fact that the human observations are also constrained by the environment, night illumination, and object and background reflectivities. At this stage the reader is exposed to a discussion as to how these limitations can be overcome to a reasonable extent by optical considerations and by technological developments in photocathodes, phosphors, fibre optics, and electron optics. This study also helps the reader to familiarise himself in depth with the evolution of the image intensifier tubes and their utilization in instruments of military interest. The text is an effort to consolidate basic as well as technical information directly related to night vision based on image intensification in precise and concise manner within the confines of a single volume and it is well done. The monograph has been well supported by an exhaustive list of references. A number of these references are also intended to help an interested reader to probe into the independent technologies which amalgamated to result in night vision.

(Dr APJ Abdul Kalam)

PREFACE Vision during the night has been one of the interesting ambitions of the humankind and for quite sometime it was considered to be within the realm of unattainability. Yet in the early twentieth century the scientific community did think of its possibilities. The importance of light-gathering by a relatively aberration-free optical system was well realized. In fact a 750 binocular with an aperture of 50 mm and an exit pupil of 7 mm to match the human scotopic eye-pupil size was referred to as a nightvision binocular in earlier literature. These did perform well at dusk and dawn though not during the night and helped in early morning assaults by an infantry column. Modern interest in the field arose with an explanation of the photoelectric effect by Einstein in 1905, discovered earlier by Hertz in 1887. Though attempts to develop suitable photocathodes based on photoelectric effect for image intensification began in early 1930s, yet the first success with an instrument system fabricated around near-infrared sensitive AgO-Cs photocathode, came in 1950s only. This resulted in ‘O’ Generation night vision instrumentation wherein the night scene had to be irradiated with near infrared radiation cutting out the visible and the reflections thereof made visible on the phosphor screen of an image converter tube. The tube itself was the result of a composite effort which brought together the photocathode technology, electron-optics for the amplification of weak electrons through an electro-optical system and phosphor screen development, all in a vacuum envelope besides suitable entrance and exit surfaces. It was by now clear that what is required is the development of better and better photocathodes corresponding to the natural illumination in the night-sky, better methods of amplification of energy and number of weak electrons and more suitable phosphors for ultimate viewing besides suitable input and output windows for the vacuum tube that may ultimately be designed. This development was further accelerated as by then the upcoming television industry was also looking for suitable phosphors and photocathodes to suit their requirements. It was hence logical that the next generation of instruments developed for night vision were passive in nature, i.e., where imaging was based on the night sky illumination itself thus dispensing with any artificial irradiation.

(xii)

Generation I image intensifier tubes were the first to appear which involved a major contribution in terms of its fibre optics input and output windows and a photocathode much more sensitive to the overall spectral distribution of the night sky. The earlier photocathodes had a lower quantum efficiency and hence three such tubes had to be coupled to give an adequate light amplification for vision without losing on the resolution. It was only a matter of time before Generation II image intensifier tubes appeared with photocathodes on the military scene by introducing electron energy amplification and electron multiplication through microchannel plates (hollow-fibre matrix) to enable adequate light amplification to be achieved with only one tube. Developments have since continued both in evolving more and more sensitive photocathodes and better and better designs for the microchannel plates. Understanding of the functioning of a photocathode resulted in the evolution of modern day negative electron affinity photocathodes. Nevertheless it can be stated that scope still exists in engineering newer and newer photocathodes with still higher quantum detection efficiencies with matching electron-optics and microchannel plates with better and better signal-to-noise ratio. This monograph has been organised in nine chapters. The first chapter on Vision and the Human Eye discusses the background against which all vision including night-vision instrumentation has to be ultimately assessed. The next chapter Search and Acquisition relates to the parameters that contribute towards establishing a visual line to an object of interest. The criteria for detection, orientation, recognition and identification are examined as also the relationship of contrast to help search and acquisition. Chapter III discusses the environment that is mainly the atmosphere (intervening medium), its attenuation of the optical signal and thereby the effect on contrast and visibility. The next chapter examines night-illumination in detail as also reflectance from surfaces of interest and from the background. After familiarizing with all the factors that affect instrument design for night vision applications, it is but natural to consider various design aspects such as those related to optical parameters, the evolution of photocathodes and the development of phosphors before one goes into the details of the image intensifier tubes which form the mainstay of night vision systems based on image intensification. Chapters V, VI, VII are therefore devoted to each of the factors, i.e., Optics, Photocathodes and Phosphors. Chapter VIII on Image Intensifier Tubes includes discussion on electron-optics and fibre optics that is relevant to the making of intensifier tubes. Chapter IX then concludes by drawing attention to overall

(xiii)

considerations for instrument design for night vision systems. Photographs and illustrations of some interesting systems designed and developed by Defence R&D Organisation also find a place in this final chapter. This monograph is limited to night vision based on image intensification. Though references in the text to ‘thermal imaging’ do find a place here and there, this text does not include the contemporary development in night vision based on thermal imaging. Obviously that should form the subject matter of an independent volume.

Dehradun

R Hradaynath Former Director & Distinguished Scientist Instruments R&D Establishment DR&DO, Dehradun

ACKNOWLEDGEMENTS The author’s thanks are primarily due to Dr APJ Abdul Kalam who initiated this idea of a monograph on night vision technology. Many thanks are certainly due to Dr SS Murthy, former Director, DESIDOC, and the present Director Dr Mohinder Singh for their persistence and patience and to the group of scientists who helped me in literature search and in consolidating the contents of this volume. Thanks are also due to a few scientists at IRDE, Dehradun, who helped me in obtaining some specific literature and by way of discussions. Particularly, my thanks are due to Shri E David who helped me with various figures and photographs that have been included herein. I am also indebted to Shri M Srinivasan of BeDelft, Pune, for the photographs referred to in Chapter VIII, and to the Director, IRDE, for all the other photographs. Finally, I like to record my sincere thanks to Shri KK Vohra who provided me with working space and to Shri Swaroop Chand an Ex-Soldier working with Shri Vohra, for his diligent day-to-day assistance.

R Hradaynath

CHAPTER 1 VISION & HUMAN EYE 1.1

INTRODUCTION Vision entails perception—by the eye-brain system, of the environment based on reflectance of the static or changing observable scene illuminated by a light source or a number of sources, and that of the sources themselves. In most cases, the illumination is natural and due to sun, moon and stars along with possible reflectance of these sources by clouds, sky, or any land or water mass. These days artificial illumination is also of significance. The ability of a living species to recognize and represent sources, objects, their location, shape, size, colour, shading, movement and other characteristics relevant to its planning of action or interaction defines its observable scene. The observable scene would thus be limited by the capability of a species and the information sought by it. Sustained vision would further require large steady-state sensitivity to properly react to amplitude and wavelength changes in the illuminating sources. Thus perception of a given scene should not get distorted by observation from sunlight at noontime to starlight at night, or under a wide range of coloured or white artificial sources or by facing away or towards the sun. Vision as perceived above would therefore call for processing of the input visual signal to attain what has been stated. For instance location of objects in space or their movement may be helped by (a) Stereopsis, i.e., using cues provided by the visual input in two spatially separated eyes. (b) Optic flow, i.e., by using information provided to the eye from moment to moment (i.e., separated in time) (c) Accommodation, i.e., by determining the focal length which will best bring an object into focus, and

2

An Introduction to Night Vision Technology

(d) Segmentation, i.e., the process of extracting information about areas of the image (called regions or segments) that are visually distinct from one another and are continuous in some feature, such as colour, depth or motion. As these are processes that can take place all over the image, parallel processing by the visual system would be quite in order. Likewise, the variable reflected optical signal received by the eye is processed by the visual system over a wide range for constancy of luminance, colour and contrast by appropriate networking of the individual signals from each photoreceptor. However, recognition of a source either by direct viewing or by specular reflection would need a different type of processing for its brightness. 1.2

OPTICAL PARAMETERS OF HUMAN EYE This monograph is restricted to the optical and processing aspects of the human eye and retina, though in actual practice the entire biological processes of the eye-retina-brain combination needs to be discussed and understood as far as presently known. It was Helmholtz[1] who suggested a schematic eye which is a close representation of the living eye with fairly accurate characteristics as defined by the first-order theory of geometric optics. Figure 1.1 shows a cross-section through such an eye 22.38 20 2.38 1.96 15 R1

7.38 H

F

I

N NI R2

R3

FI

H 6.96 3.6 7.2 Figure 1.1. Optical constants of Helmholtz's schematic eye (all dimensions are in mm).

Vision & human eye

while Table 1.1 details its optical parameters[1,2,3]. Depending on the degree of accommodation desired, the radius of the anterior lens surface is assumed to change up to + 6.0 mm, while everything else remains fixed. The cornea is assumed to be thin, and so also the iris which is supported on the anterior lens surface. The optical parametric values are for sodium light. Table 1.1. Parameter

Optical parameters of the schematic eye

Symbol (see Fig. 1.1)

Radius (mm)

R1

8

0

41.6

R2

10

3.6

12.3

—

F H

— —

–13.04 1.96

— —

— —

N

—

6.96

—

—

R3

–6

7.2

20.5

—

F

H

— —

22.38 2.38

— —

— —

N

—

7.38

—

—

Entrance — pupil position Exit pupil — position

—

Cornea

Anterior Lens surface Focal plane Principal plane Nodal plane Posterior Lens surface Focal plane Principal plane Nodal plane

Volumes Eye lens Anterior chamber Posterior chamber Eye as a whole

Distance from corneal vertex (mm)

Refractive power (diopters)

Refractive index

(for practical purposes a thin curved surface)

3.04 — (size 1.15  pupil diameter) — 3.72 — (size 1.05  pupil diameter)

—

— —

— —

— —

30.5 —

1.45 1.33

—

—

—

—

1.33

—

—

—

66.6

—

—

The aberrations of the eye are well documented elsewhere as also the errors of refraction. Methods do exist for evaluating the line spread function of the eye, retina and the entire visual system experimentally, as also some of its geometric aberrations. It is

3

4

An Introduction to Night Vision Technology

interesting to note that the eye focused for infinity exhibits positive spherical aberration and for very near distances negative, while for intermediate distances (around 50 cm) it is essentially zero. The line spread is minimum for a pupil diameter of 2.4 mm, and for smaller diameters, the spread approaches the diffraction limit. At 2.4 mm also it is almost diffraction limited with an exponential fallout representing scatter and depth of focus. As the pupil diameter increases beyond 2.4 mm, the fallout becomes more prominent and dominates the Guassian spread. Figure 1.2 is basically a sketch showing the blood supply to the eye representing arteries and veins as shaded and dark lines, respectively [4]. The cornea (C ) with the sclera (S ) represent the outer fibrous envelope of the eyeball. While the cornea is transparent, the sclera is pearly white. The sclera is almost five-sixths of the envelope. The two structures are dovetailed into one another biologically. The cornea is thickest (about 1mm) posteriorly, C

I L S Ch R

N

Figure 1.2. Half cross-section of the representative biological eye (C – Cornea, S – Sclera, Ch – Choroid, R – Retina, N – Optic Nerve, I – Iris, L – Lens. The black and shaded parts denote veins and arteries).

Vision & human eye

gradually becoming thinner anteriorly. At the site of the optic nerve, the sclera splits up into a network of interlacing bundles, called the lamina cribrosa, leaving a series of fine sieve-like apertures, through which the bundle of optic nerve passes from the eye. The choroid (Ch) is interposed between the sclera (S ) and retina (R) and is chiefly concerned along with the ciliary body and iris (I ) in supplying nutrition to the internal parts of the eye. It forms a continuous deeply pigmented coat, except at the entrance of the nerve into the globe. Nourishment to the retinal pigment layer and the outer retinal layers is provided by the choroidal capillaries, while the innermost layers of retina are served by the retinal artery. The retina (R) is a membrane containing the terminal parts of the optic nerve fibres, supported by a connecting network. It lies between the choroid and the membrane enclosing the vitreous body. It diminishes in thickness from 0.4 mm around the optic nerve entrance to 0.2 mm towards the frontal side. It is perfectly transparent and of a purplish red colour, due to the visual purple present in the rods. Viewing from the front there is a yellowish spot somewhat oval in shape with its horizontal axis measuring around 2-3 mm. A small circular depression in the centre known as fovea centralis has the maximum packing of cones in an area around 3 mm in diameter. Corresponding to the entrance of the optic nerve, one observes a whitish circular disc of around 1.5 mm diameter known as the optical disc, which presents a blind spot as this area has no nerve-endings. It is about 3 mm to the nasal side of the yellowish spot. The light entering the cornea passes through the full thickness of the retina which is a thin 350 m sheet of transparent tissue and optical nerve head to reach the layers of rods and cones. The properties of rods and cones are very vital as photoreceptors. It is well known that cones in the macular region, i.e., fovea centralis and its surround are highly packed at around 0.003 mm centre to centre. Appreciation of form and colour therefore is better possible with cone-vision which responds above a certain visual threshold. At the same time it is known that our rods are capable of signalling even the absorption of single photon and signal low light level phenomena though without clear appreciation of form and colour. Electric impulses arising in these photoreceptors are transmitted via retinal interneurons to the innermost ganglion cell layers with around 100,000 cells. The axons of these cells form the

5

6

An Introduction to Night Vision Technology

optic nerve and convey the information further to the various areas in the visual system of the brain. The iris (I ) arises from the anterior surface of the ciliary body and results in an adjustable diaphragm the central aperture of which is known as pupil. The diaphragm divides the space between the cornea and the lens into two chambers which are filled with a fluid – the aqueous humour. The ciliary body is in turn, a continuation of the retina and the choroid. The iris has a firm support by lying on the lens. The contractile diaphragm reacts to the intensity of light and accordingly adjusts the pupil diameter from 7 mm to 2 mm from starlight to noonlight. In a given position it also cuts off marginal rays – which unless stopped would diminish the sharpness of the retinal image. The lens (L) is a transparent, colourless structure of the lenticular shape, of soft consistence enclosed in a tight elastic membrane whose thickness varies in different parts of the lens. The circumference is circular, 9 mm in dia, with the central thickness as 5 mm in an adult. The posterior surface is more highly curved, and embedded in a shallow depression in the vitreous humour, while the anterior surface is in contact with aqueous humour. The vitreous humour is a transparent colourless gelatinous mass which fills the posterior cavity of the eye and occupies about four-fifths of the interior of the eyeball. The aqueous humour is transparent and colourless fluid and serves as a medium in which iris can operate freely. The optic nerve (N ) collects its fibres from the ganglion in the retina and passes through the eyeball. The fibres from the right halves of both the retinas pass into the right optical tract and the fibres from the left halves pass into the left optical tract, each tract containing a nearly common field of vision from both the eyes. Both the tracts continue to the centre of vision in the brain. 1.3

INFORMATION PROCESSING BY VISUAL SYSTEM

The continuous photon stream that is incident on both the eyes as a result of light reflectance from the environment is appropriately focused on the retinal receptors (i.e., rods and cones) through its optical system (i.e., cornea, pupil, lens and the intervening spaces occupied by aqueous and vitreous humour). This photon stream variable in space (x,y,z), time (t) and wavelength () is sampled in space and wavelength by the three types of cone

Vision & human eye

LOG RELATIVE SENSITIVITY

receptors each sensitive to red, green and blue and appropriately filtered by the spatial and chromatic apertures of these receptors. See Fig. 1.3 for their response.

RED

BLUE

400

600 500 WAVELENGTH (nm)

GREEN 700

Figure 1.3. Relative spectral sensitivities of the three types of cones; blue, green and red (determined against a dim w hite background).

Simple sums and differences of these signals results in an achromatic signal, a red-green signal, and a blue-green signal. A parallel partition of the image is by spatial location, approximately the bandwidth of visual neurons. The phsycophysical and physiological evidence to date suggest a partition in just two temporal bands. The temporal frequency processing may be in terms of static and moving components. The role of retina as a processor is significantly complex. Apparently immediately after the first layer of receptors, every neuron receives multiple connections from multiple varieties of adjacent interneurons with various anatomical and physiological roles. Each neuron has typically 50,000 connections with other neurons. Thus even before the signal leaves the retina, each ganglion (in the final neural layer) relays information as based on interactions between several receptor types, non-linear spatial summation, both

7

8

An Introduction to Night Vision Technology

subtractive and multiplicative inhibition, far field effects, and so on. At the stage of the ganglion layer in the retina, it is observed that this layer contains two intermixed cell types that differ in size and in the way they integrate the signals from the cones and the rods via the inter-neurons. The final signals from the retina itself are thus in two parallel channels which reach the identified regions, i.e., the magno- and parvocellular systems of the lateral geniculate nucleus of the thalamus[5]. Each system is identified with certain vision parameters. Thus, the parvocellular system has information regarding colour, high spatial resolution, low contrast sensitivity and is slow (static sensitive). The magnocellular system on the other hand is colour blind, has high contrast sensitivity and low spatial resolution, carries stereopsis information, and is fast (movement sensitive). This information which goes to different layers of the primary visual cortex is analysed for the perception of (i) colour (ii) form, and (iii) movement, location and spatial organisation in at least three separate processing systems. Thus the two pathways at the lower level seem to be rearranged into three subdivisions for relaying the information to higher visual areas. That movement, high contrast sensitivity and stereopsis information is carried through one system and leads one to predict that movement of either the observer or the object, or stereo-viewing two images of the same object would lead to easier detection of hard-to-see objects. Vision, is served both by the top-down and bottom-up processing procedures, possibly simultaneously. The bottom-up procedure analyses the stimulus in terms of the information sought by the retinal processes as described above, processing parallely small elements of a scene, then joining them into larger and larger groups, and ultimately presenting it as a single scene. The parameters of analysis could be uniformity of shading or colour and their variation, certain geometries like edges, convexities, etc., (segmentation), or movement (optic flow), or depth (stereopsis) and orientation. The top-down principle would operate via organised percepts recorded in our memory and improvements on it in case something new has been observed which was not in the memorypackage earlier. The bottom-up process can be thought of as datadriven while the top-down could be referred to as knowledge driven. A considerable amount of the bottom-up processing is done in the retina itself. Overall mechanisms could be discussed as under, in terms of response to light stimulus, contrast and colour.

Vision & human eye

1.4

OVERALL MECHANISMS

1.4.1

Light Stimulus Light stimulus experienced subjectively as brightness is measured in terms of units of luminance, i.e., candela per square metre (cd/m2). Based on this, one can define the luminosity function of a standardized eye for cones and rods independently, i.e., for photopic and scotopic vision. Figure 1.4 shows the normalised relative spectral sensitivity. However, it is important to take into account the pupil size and the luminosity function of both cones and rods, so that the retinal illumination is correctly measured.

RELATIVE SENSITIVITY

1.00 0.80

SCOTOPIC

PHOTOPIC

0.60 0.40 0.20 0.00 400

450

500

550

600

650

700

WAVELENGTH (nm) Figure 1.4. Normalised spectral sensitivity of luminosity functions for scotopic and photopic vision.

The unit of retinal illuminance troland (td) is a multiple of the luminance (L) in cd/m2 and pupil area (P ) in mm. As the luminosity functions of both scotopic and photopic vision are different, the scotopic and photopic trolands get a different value (Fig. 1.5). The troland values can be converted to photon values. The determination of the actual retinal illumination in terms of absorbed photons requires assumptions about transmission losses from the corneal surface to the retina caused by the entire optical system of the eye, as also by the probability of photon absorption particularly at very low light levels. According to a number of workers, in scotopic vision the number of photons that excite the rods is

9

10

An Introduction to Night Vision Technology Luminance (log cd/cm2)

Pupil diameter (mm)

–6

–4

–2

7.1

6.6

5.5

0

4.0 1.1

–4.0 scotopic

–2.1

–0.22

starlight

Luminance of white paper in Visual function

2.4 rod

photopic

Retinal luminance (log td)

2

scotopic scotopic threshold

6

8

2.0

2.0

2.0

cone 4.5

6.5

8.5

0.70 moonlight indoor lighting

mesopic photopic threshold

no color vision, poor acuity

2.6

4

sunlight

photopic best rod damage acuity saturation possible begins good colour vision, good acuity

Figure 1.5. Relationship of luminance, pupil diameter and visual function. Dashed curves represent measured stimulus response in single primate rod and cone.

around 25 per cent of those incident on the cornea, though the signal transmitted to the brain suggests the arrival of around 5 per cent. Yet rods have been shown to be highly sensitive and even signalling the absorption of single photons. Experimental estimates show that one scotopic troland corresponds to about four effective photon absorptions per second. The relationship of luminance, pupil diameter, and retinal luminance over the entire visible range is shown in Fig. 1.5. Threshold vs Intensity Functions & Contrast Threshold vs intensity function in respect of photopic (cone) vision referred to as the Weber-Fechner function is shown in Fig. 1.6. The function can be put down to a fair approximation in the form: 1.4.2

L / L0  k ( L  L0 ) n

(1.1)

Where L is the incremental luminance as a function of the luminance background L, and L0 is the threshold luminance increment for a dark luminance value L0 which is just at the threshold vision of the eye. Power n has a value from 0.5 to 1 while k is a constant. When L>>L0 i.e., beyond a certain value of retinal luminance and n = 1, L/L is a constant. The constancy of L/L

Vision & human eye 4

LOG ( L /  L o )

3 L

2

L Lo

1 n =1 0

Lo 

-1

0 1 LOG L (td)

2

3

4

Figure 1.6. Threshold vs intensity function in respect of photopic vision.

known as Weber’s Law explains that contrast remains constant with changes of luminance in an observable scene above a certain minimal level of retinal luminance. It has to be noted that Weber’s Law is operative for luminances as reflected from the observable scene and analysed by the eye-brain system. The perception of light sources and brightness as such is in addition to the operation of the Weber’s Law. Various definitions of contrast are in use, but as all these are based on a ratio, they all yield invariance of contrast from a change in illumination level. A threshold-vs-intensity curve for rod vision is shown in Fig. 1.7. It can be observed that rods unlike cones are saturated by steady backgrounds as a consequence of their high sensitivity and lack of gain control. At the lower light levels, they are known to signal the arrival of even single photons. Obviously, because of their low spatial resolution and saturation, their contribution to daytime vision is rather insignificant as against cones. 1.4.3

Colour As indicated earlier, the colour sensation is picked up by three independent types of cone photoreceptors with the spectral characteristics as shown in Fig. 1.3. The three cone types are designated blue (B ), green (G ), and red (R). Constancy of colour has

11

An Introduction to Night Vision Technology

6 5 4 Log (L/Lo )

12

3 2 1 0 

-4

-3

-2

-1

0

1

2

3

LOG L (SCOTOPIC td)

Figure 1.7. Threshold vs vision.

intensity function in respect of scotopic

also to be sensed in the same way, as the constancy of contrast with changes in illumination or in its spectral content. The sensitivities of the three types of cones are adjusted in such a way that response ratio when adapted to coloured light is the same as that produced in white light. Chromatic adaptation is also necessary for correct perception of object colour. Changes in spectral illumination can really be significant as shown in the reflectances observed from a surface facing away from the sun, or directed at it (Fig. 1.8). Though perfect colour constancy may not be possible, the same is achieved over a good deal of spectral variation as indicated by colorimetric studies. Colorimetric units are represented by a vector in threedimensional colour space plot along the three axes representing the tristimulus values, that indirectly define how the light is registered by the three cone types: blue, green and red. As what is of interest is the relative amounts of color along each stimulus, it is sufficient to specify the chromaticity of a light stimulus with only two of the three chromaticity coordinates to record the chromatic information.

Vision & human eye

RELATIVE SPECTRAL POWER

150

FACING AWAY

FACING TOWARD

FROM SUN

SUN

100

50

0 300

400

500

600

700

WAVELENGTH (nm) Figure 1.8. Relative spectral response by a reflecting surface facing away and facing towards the sun.

1.5

IMPLICATIONS FOR NIGHT VISION

It is obvious that the naked eye responds better at low light levels as obtained under moonlight or starlight at pupil diameters of 5 mm and more, utilizing scotopic vision (rods as sensors). Yet as visual acuity is best with cone vision and at higher light levels, it becomes imperative to come to at least that level of luminance in the scene as is essential for desired spatial location of objects in the object-scene (Fig. 1.5). Practical studies of the line transfer function also show greater departure from the diffraction image of a line at pupil size of 4 mm and above and better matching around 2 mm. This could be achieved by either illuminating the night scene artificially or by intensifying the image of the objectscene. Both alternatives create a hybrid situation for visual adaption as the observer is in a dark area where the eye is adapted for rodvision, but looking on a scene that is either illuminated artificially (for a few seconds or minutes) or whose image is intensified. To get the best results therefore training for sustained foveal vision under overall dark conditions may be essential. As optic flow and stereopsis help better understanding, binocular observation is preferable. Earlier designs also visualised increase of instrument optic apertures in such a fashion that the illuminance at the eye pupil could be increased manifold. Such devices did play a significant

13

14

An Introduction to Night Vision Technology

role under dawn and dusk conditions in military operations. Astronomical telescopes also tend to increase their apertures to observe the faintest stars in the image plane directly or for recording on photographic plates or by utilizing image intensifiers or charge coupled devices (CCD) with appropriate sensors for the desired region of the spectrum. REFERENCES 1.

Helmholtz, H.V. "Helmholtz’s treatise on physiological optics,(Vol. 1)", Translated from Optical Society of America, 3rd ed, (German), ed. by Southhall, J.P.C. (Rochester, N.Y. 1924).

2.

Lawrance, L. Visual Optics and Sight Testing, 3rd ed. (London: School of Optics, 1926.) p.452.

3.

Walter, G.D. (Ed). The Eyes and Vision, Chapter 12, Handbook of Optics. (McGraw Hill Book Company). Forrest, J. The Recognition of Ocular Disease. (London: The Hattan Press Ltd). Livingstone; Margaret, & Hubel, D. "Segregation of Form, Color, Movement and Depth". Science, vol 240, pp. 740-49. Waldman, G. & Wootton, J. Electro-optical Systems Performance Modeling. (Artech House, 1993).

4. 5. 6.

CHAPTER 2 SEARCH & ACQUISITION 2.1

SEARCH

As is by now obvious, eye is basically par excellence, a spatial location and movement detection instrument under conditions of varying contrast, colour and resolution dependent on differing levels of illumination and their spectral content. Having observed a scene, need arises for search of objects of its interest. Thus a species, say a frog, would like to know about small organisms like the fly which it can eat and simultaneously be alert about the predators in its field of view. At a higher level, the task of a human being though similar is more elaborate. The humans also search and acquire the targets of their interest for desired interaction or avoidance. Refined search and acquisition has led to the evolution of a large number of techniques and utilization of parts of the entire electromagnetic spectrum beyond the capabilities of the human eye. As such it would be of interest to know about the search and acquisition techniques that are adopted by the human eye and by the instruments that we are dependent on. The image of a scene is stabilized on the retina by reflex movements of the eye to balance its involuntary movements of high frequency tremor, low speed drift and flicks, all of low amplitude, even when an observer is consciously trying to fixate on a given point. This is presumably necessitated as the rods and cones get desensitized if the illuminance of the light falling on them is absolutely unchanging. To make a search, the eye jumps from one fixation point to another, dwelling momentarily on each fixation point after each jump. The jump called a saccade has a definite amplitude. Search-time would be excessive if the dwell-time after each saccade is long and the saccades are small. It has been experimentally observed that if the observed sector is larger, the

16

An Introduction to Night Vision Technology

observer makes larger saccades between fixations and dwells a shorter time at each fixation[1]. Empirical formulae correlating fixation (glimpse) time and the field have been reported[2] as t  0.6836  0.2132

(2.1)

where t is the average fixation time,  is the search sector in degrees and s  0.152 t 9.127

(2.2)

where s is the average saccadic size. In another approach, both the fixation time and the saccadic size may be considered to be fixed, i.e., the fixation time as 0.3 s and saccadic size as 5° [3]. Then the time ts to completely search a sector  degrees by  degrees is given by

t s  0.3( / 5)(  / 5)

(2.3)

These approximations seem to be reasonable for search sectors from 15° to 45° [4]. This has particularly an interesting corollary in vision optics. Thus if episcopes in a battle tank which are generally of limited field of view are not properly aligned and juxtapositioned with respect to each other for the observation of a wide angle scene at a single go, i.e., with the head and eye movement at the same place, the search time is bound to be longer and quite a few visual cues are likely to be missed. This was observed to be true of earlier designs for episcopic vision which had led to battle failures. 2.2

ACQUISITION

Once a search has been completed and it is desired to acquire the target, it is found that acquisition is possible at various levels. Thus while taking an early morning walk in an open space, one may observe at a distance slight movement at first, and not be sure about the object. Once the object is a little nearer, one may be able to decide that it is a human being, and once the human being is still nearer one can recognise the face and identify the person. A similar situation arises in battlefield conditions also, wherein one may acquire some target based on its movement, or its lack of fitment in the background, but on closer observation may identify it as an object of interest and subsequently recognize it as a tank, heavy

Search & acquisition

vehicle or a light vehicle. Further closer observation would reveal the exact type of vehicle and whether it is one’s own or enemy’s. The range of acquisition is considerably increased by optical instruments during daytime while the night vision electro-optical instruments make it possible by the night. While there are many parameters in the instrument that decide the acquisition range and subsequent detection, design factors like overall magnification, field of view, contrast rendition and quick search ability are more important. It will be appreciated that a priori knowledge helps a great deal in deciding as to whether a target is to be engaged no sooner it is acquired, or one has to wait further identification. With this background it is now possible to decide on the levels of acquisition. The standard terms used are: (a)

Detection The term implies that the observer in his search has located an object of interest in the field of view of the vision system that is being used.

(b)

Recognition This would mean that the observer is able to decide on the class to which the object belongs, e.g., a tank, a heavy vehicle, a group of people and the like.

(c)

Identification At this level of acquisition, one should be able to indicate the type of the object, i.e., which type of tank, vehicle, or the number of people in a group. An important military requirement would be identification of friend or foe (IFF).

As the vision cues during daytime (like shading, colours– their hues and variation, better sensitivity to stereopsis and optics flow) are far more except in extreme bad weather and fog, the levels of acquisition do not have that relevance as during night or when a visual scene is displayed on a monitor. The definitions of the levels of acquisition are therefore more concerned with image intensification or thermal mapping. We will therefore first summarise parameters that lead to acquisition. Acquisition has been variously discussed and found broadly to depend on search time, fixation time, and the type of vision instrumentation used and their characteristics. Models that have been used for detection or acquisition proceed on the premise that the signal strength from

17

18

An Introduction to Night Vision Technology

target vis-a-vis its background should equal or exceed the detection limits of the eye for a specified contrast ratio between target and background with variations in size, shape and luminance. 2.3

BLACKWELL’S APPROACH

Blackwell (1946) conducted an elaborate set of experiments to determine the detection capability of the humans as a function of size, contrast, and luminance. Around 4,50,000 observations made by small groups out of 19 young observers of circular targets against uniform backgrounds, both light on dark and dark on bright mostly with exposure time of 6 seconds were statistically analysed. The results were expressed in terms of contrast threshold that was necessary for a detection probability of 50 per cent. The following conclusions can be drawn from the data for near foveal vision randomly at any point 3° off from the vision axis for an exposure time of 6 seconds (i.e., involving small search)[5]. (a) (b)

The contrast threshold values decreased with increase in target size or target brightness. The contrast threshold values do not change appreciably at larger target angular size and at higher illuminance levels. Analysis showed that foveal vision was used at high brightness and parafoveal at low brightness. These conclusions can also be drawn from the later data presented in 1969 by Blackwell and Taylor[6] with regard to direct line of vision detection for exposure timings of a one-third of a second and involving no search. Obviously, the first set of experiments compiled the data due to excitation of the fovea as well as a limited area of the foveal region (less than a degree). Experimental data beyond 3° seems to be limited. The Blackwell data was best fitted later into empirical equations by Waldman, et al [7] as





Log C t = 0.075  1.48  10  4   log L  2.96  1.025 log  

2

(2.4)

0.601   1.04 for high light level region (photopic vision). Log C t = 0.1047 log L  1.991  1.823  1.64 log  2

(2.5)

for low light level region (scotopic vision). The two regions of light-levels were divided at about 710–4 foot lambert.

Search & acquisition

Where Ct is contrast at threshold, L is luminance of the target against a uniform background (f L), and  is the target size in angular units (minutes). Work has continued to improve on these empirical equations to include parameters like wider fields (beyond 3°) and noise as important inputs[8]. This approach enables simulation of electro-optical systems to a better degree of predictability vis–a–vis actual field performance. The model designers have to combine detection with search based on a great deal of empirical data on contrast thresholds, response times, fixation times, saccadic size etc, for a specified target and its backgrounds besides incorporating the human factors and parameters of instrument design that are inherent in observation. Though many models have been developed using modern computers, the predictability has yet to reach a standard for reliability. Research based on the above approach leads to modelling for acquisition and also suggests that in experimentation an equivalent disc object could also reproduce the behaviour of a given object of military interest in the field. There is experimental evidence to support that detection of smaller and smaller discs could simulate for observing the details of a particular object, i.e., simulate for detection, recognition and identification particularly for imaging through intensifiers or thermal imagers. Also where one is aware of the direction of likely appearance of the object, the aim may be restricted to know whether this is an object of interest. Air Standardization Agreement of 1976 sets minimum ground object sizes required in imagery for various levels of acquisition of various targets. 2.4

JOHNSON CRITERIA

Following a suggestion in early 50s from Coleman that one might establish a relationship between a real target against a background and a target made of contrasting line-pairs, Johnson (1958) decided to extend the spatial frequency approach to night vision systems[9]. His approach involved classification of models of objects of interest based on their silhouettes, shape and equivalent area blobs (for detection) set alongside a bar-test pattern with matching contrast and observing these from a given distance, as the illumination in the test area was increased from zero. It was possible to conclude based on a data of some 20,000 observations

19

20

An Introduction to Night Vision Technology

that a relationship existed between the number of lines resolved at the target and the corresponding decisions of detection, recognition and identification. The targets tested were truck, M-48 tank, Stalin tank, Centurian tank, half truck, jeep, command car, standing soldier, and 105 Howitzer. It was recorded that there was a certain uniformity in observation regarding line-pairs for a critical target dimension. The following data was derived for resolution per minimum dimension across the complete object[10]: • • • •

Detection has an average of 1.0 ± 0.25 line-pairs Orientation has an average of 1.4 ± 0.35 line-pairs Recognition has an average of 4.0 ± 0.8 line-pairs and, Identification has an average of 6.4 ± 1.5 line-pairs Figure 2.1 illustrates the Johnson approach schematically.

The relationship can be extended to TV lines per minimum object dimension where detection, recognition and identification would have 2, 8, and 12.8 TV lines respectively. Johnson criteria could be easily appreciated both in the laboratory and in the field and hence has been very widely used all over the world, not withstanding the fact that these criteria were not completely applicable to different viewing angles and had been based on threshold of vision and did not allow for calculations for range of different probabilities. This approach lends itself to easy interpretation in the spatial domain, where when an optical chain is understood to be a linear system, the system performance can easily be predicted based on the MTF values of the objective, image intensifier and its subunits, and lend the answer in terms of linepairs per mm for the entire system for any given contrast value. Using these criteria, search effectiveness was also tried to arrive at a trade-off between gain, contrast and resolution on response times in search operations through a night vision system. This work showed that gain was much more significant for search than for static viewing. 2.5

DISPLAY SIGNAL-TO-NOISE RATIO

As a good part of the night vision imagery is through a video display, one may emphasize signal-to-noise ratio in the signal to the eye from the display. To assess this aspect it may be assumed that the resolution of the system as a whole is such that in an equivalent direct vision system, the object of interest could have been easily resolved. Next, the object of interest could be thought of

Search & acquisition

Figure 2.1. Equivalent bar targets for various field targets

21

22

An Introduction to Night Vision Technology

as an isolated rectangular target or to be in more consonance with the Johnson criteria as a periodic bar chart[11]. The signal-to-noise ratio (SNRDi ) in the two cases could be defined as

SNRDi  [ 2 tf ( a / A)]1/ 2 SNRv where

(2.6)

t

= integration time of the eye f = video bandwidth  a = area of the target in the image A = area of the field of view SNRv = signal-to-noise ratio in the video signal and for a periodic target SNR Di  2tδ f /α 

1/ 2

U / N 1 / 2 SNR v

(2.7)

where = displayed horizontal-to-vertical ratio = bar length-to-width ratio = bar pattern spatial frequency (lines/picture height) These expressions in a realistic case could be modified by involving the MTF of the system. The point to note is that detection experiments provided a value of around 3 for the SNRDi for the threshold value at 50 per cent probability. The value appeared to vary only slightly for a wide range of rectangular shapes and sizes and also for squares. The periodic bar target showed a variation for both spatial frequency and length-to-width ratios of the patterns. Further experiments in recognition and identification of military vehicles followed theoretical calculations for the SNRDi for Johnson’s equivalent bar pattern and indicated the value as 3.3 to 5.0 for recognition and 5.2 to 6.8 for identification against various backgrounds. The variability of values to such a large extent suggests that this parameter is not a very good general performance measure though one could draw on the minimum values that may be necessary for good performance.  U N

2.6

DETECTION WITH TARGET MOVEMENT Detection of movement is a parameter of importance both by day and night, and the expected sensitivity to movement is an inbuilt faculty of the human eye-brain system. In field conditions this would imply our sensitivity to the angular movement of a likely target or object of interest. While the detection probability is enhanced, visual acuity would drop, i.e., perception or detection

Search & acquisition

would be much easier while recognition and identification much more difficult. Experiments have led to the following empirical equation which shows the relationship of the contrast of a moving object (Cm ) to that of the same object when stationary [12]: C m =C (1+0.45 w 2 )

(2.8)

where C= w= 2.7

Contrast while stationary Target angular speed in degrees per second for speeds up to 5° per second.

PROBABILITIES OF ACQUISITION

Detection probability can be calculated as a product of a number of probabilities based on the variables in the observer – object of interest scene. The factors could be the evaluation of a target embedded in its background, intervening atmosphere, clutter, obscuration, the capabilities of the electro-optical system used, and the display parameters. In addition, human factors, such as training, search, and establishing a line of sight between the target and the sensor also matter. Further recognition and identification could be done by involving Johnson criteria. All these parameters and possibly more have been selectively incorporated in various models with appropriate algorithms to arrive at a possible prediction of the field conditions. The approach has been of interest to many a workers, and models have been developed based on image intensifier and forward looking infrared systems. While a universal model is far from developed, one can possibly select a model for advance understanding limited to certain parameters, such as atmospherics or variations in instrument design. It is interesting to note that this approach is rather late in the day for most of the systems already developed, but may have a significance in sophisticated futuristic developments. 2.8

CONTRAST & ACQUISITION

It is now obvious that consideration of contrast really leads to the detection and acquisition of an object of interest. It may therefore be worthwhile to directly interpret acquisition in terms of contrast at the object and its transfer to the contrast in the image as seen by the eye. The image contrast factor Ci would be dependent on, (if the object structure is to be perceptible) the ratio between the random fluctuations n during the observation time to the mean number of quanta received by the eye n~ [13]:

23

24

An Introduction to Night Vision Technology

C i

 (n )

(2.9)

n~

The factor of perceptibility Kr would depend on the perception of the object and be a common value for simpler structure like rasters, bar-charts, Landolt’s rings, disc objects and the like and would be the constant of proportionality, i.e.,

Ci K r

 (n ) n~

(2.10)

n , while The structural content is carried in the factor (n)/ ~ the perception, a result of eye-brain combination in factor Kr . Obviously, in a real situation one would like to find out the relationship of the perception with the object contrast, Co . It is not difficult to do so, if one knows the modulation transfer function for a given spatial frequency a and defines it as Tm (a) for the complete electro-optical system. The equation can be put down as

Co 

Kr  na  , ~ Tm a  n a 

(2.11)

where  (n) (a) is the random fluctuation and n~ (a) is mean number of quanta received by the eye during the observation time at a spatial frequency (a). Apparently, the total perceptibility would be a summation of all such contrast values at all spatial frequencies of our interest. The structural content of the image at the retina, i.e.,  (n) / n~ would arise as a result of the total system noise-to-signal ratio and could best be assessed at the retina itself, if it were possible. For practical purposes, it can be approximated most closely by measuring the output signal-to-noise ratio of an imaging system when its gain is such that it makes only the noise detectable. Variants of this measure, in terms of equivalent background illumination, background dark current, or the noise equivalent power are different definitions in varying context for different detectors to arrive at the same parameter which hopefully would help us in predetermining what we are looking for. One can argue from the fundamental reasoning that the lowest possible noise-to-signal ratios could be achieved if every

Search & acquisition

quantum carrying some structural information about the object could be processed by the sensor system so that ultimately each such quantum could give an identifiable signal. In such a case one could define the quantum detection efficiency qde by a factor F as 2

F 

  n    ~   n   out 2

  n    ~   n   in



1 qde

(2.12)

It is a highly remarkable property of our visual system that the same relationship holds true for perceptibility experiments irrespective of the nature of quanta, e.g., for x-ray photons. At retinal level in each case, it is conversion of visible photons into discrete nerve signals that leads to perception. If  (n) tends to become larger in comparison to ~ n , i.e., the fluctuations are far in excess to that of the mean number of photons received by the eye, it is the fluctuations which become dominant and the structural content gets suppressed. Experiments show that if this factor is more than one-third, the relative fluctuations become perceptible destroying the structural content of the object in a wide range of luminance. Considering Eqns (2.11) and (2.12) one could work out the object contrast in terms of the quantum detection efficiency as under, summing up overall the spatial frequencies of interest:

Co 

Kr F   n   Tm  ~   n  in

(2.13)

This approach has the advantage that when one limits to simpler structures like bars, line, raster, discs or the like usually used as test objects, the perceptibility factor can be treated as a constant and contrast at the object conveniently evaluated from modulation transfer function of the system, qde and the input noiseto-signal ratio. If one were to critically analyse the definitions and terms used both in the visible and the infrared regions even at low light

25

26

An Introduction to Night Vision Technology

levels, one would find that the attempt is to get at the ultimate perception through all the parameters that go to define a sensor and a complete system. That it has not been completely done and is still a matter of research is evident in the fact that the success or the failure of a given system with a given sensor under actual field conditions cannot be predicted accurately. The structural factor  (n ) / n~ that is accepted by the system is also fluctuating as a whole at the entrance aperture. One can normally use qde to define the performance of imaging sensors in general—be it photodetectors, image tubes or the like. In the case of photopic imaging the number of quanta available is so large that statistical fluctuations are negligible and ~ as a constant. The imaging one can treat the factor  (n ) / n performance thus depends upon modulation transfer function, Tm . The qde is not of much concern in direct vision due to the abundance of quanta with structural information. In case of low light level imaging, the number of quanta available is so small that performance is governed by statistical fluctuations. While  (n ) / n~ determines the image contrast, the improvement in MTF of the system is not going to play that vital a role as in the case of normal day instruments. In case of x-ray, ultrasound or NMR imaging, there are undoubtedly fundamental statistical limits but the method of calculating these limits is not obvious. Due to various other phenomena associated like scattering, differential absorption etc., the value becomes so much complicated that the theoretical prediction is difficult and one resorts to signal-to-noise ratio. In these cases also, MTF (Tm) is not very important and one works out such image processing methods so that the contrast rendition is improved to detect contrast lower to 0.001 per cent although resolution may be as low as a few lines/mm. The goal of being able to thwart the natural conditions which limit the usefulness of both vision and photography in extracting information from scenes of decreasing low apparent object contrast has been pursued with only limited success for many years. The use of short wavelength cutoff filters, for example, combined with the long wavelength end extension of photographic film sensitivity has helped to penetrate the veil in cases where contrast has been reduced to wavelength-dependent (Rayleigh and Mie) scattering. The introduction of infrared sensitive emulsions carried this photographic approach as much as it could.

Search & acquisition

Other types of imaging systems operating both in the visible and near-infrared portions of the spectrum (classified as electro-optical imaging systems) and relatively long wavelength forward-looking infrared (FLIR) sensor systems have been developed which can detect, quantify and display significantly lower contrasts than those possible with conventional photography. The common properties of these systems which allow them to perform at low contrast include wide dynamic range and a high level of linearity. These properties in turn allow for subtraction of an absolute dc background level; effectively ac coupling carried to an ultimate extent. Signal-to-noise for the system as a whole thus becomes the most relevant parameter.

REFERENCES 1.

Enoch, J.P, "Effects of the Size of a Complex Display Upon Visual Research". J. Opt. Soc. Am. vol 49, (1959), pp. 280-86.

2.

Waldman, G.; Wootton, J.; Hobson, G. & Luetkemeyer, K, "A Normalised Clutter Measure for Images". Comp. Vis., Graphics and Image Pro. vol. 42, (1988), pp. 137-156.

3.

RCA Electro-optics Handbook. Tech Series, EOH –11, (RCA Solid States Division, 1974).

4.

Waldman, G. & Wootton, J, Electro-optical Systems Performance Modeling . (Artech House, 1993).

5.

Blackwell, H.R. "Contrast Threshold of the Human Eye". J. Opt. Soc. Am. vol. 36, no.11, (1946), pp. 624-43.

6.

Blackwell, H.R. & Taylor, J.R, Survey Of Laboratory Studies of Visual Detection. NATO Seminar on Detection, Recognition. and Identification of Line of Sight Targets. (The Hague, Netherlands, 1969).

7.

Waldman, G.; Wootton, J. & Hobson, G, "Visual Detection with Search: An Empirical Model". IEEE Trans. on Systems, Man & Cyber. vol. 21, (1991), pp. 596-606.

8.

Overington, J. "Interaction of Vision with Optical Aids". J. Opt. Soc. Am., vol. 63, (1973), no.9, pp. 1043-49.

9.

Johnson, J, "Analysis of Image Forming Systems". Image Intensifier Symposium. (Fort Belvoir VA . October 1958).

10.

Wiseman, R.S. "Birth and Evolution of Visionics". SPIE, Infrared Imaging. vol. 1689, (1997), pp. 66-74.

27

28

An Introduction to Night Vision Technology

11.

Rosell, F.A. & Wilson, R.H, Recent Psychophysical Experiments and the Display Signal-to-noise Rate Concept. Perception of Displayed Information. (New York: Plenum Press, 1973).

12.

Petersen, H.E. & D.J. Dugas. "The Relative Importance of Contrast and Motion in Visual Detection". Human Factors, vol. 14, (1972), pp. 207-16. Hradaynath, R. "Opto-electronic Imaging: The State of the Art". in Proceedings of the International Symposium on Opto-electronic Imaging, (New Delhi: Tata-McGraw Hill Publishing Co. Ltd., 1985) pp. 19-33.

13.

CHAPTER 3 THE ENVIRONMENT 3.1

INTRODUCTION

The environment has an important effect on the observation of a target or an object of interest; the most important single parameter being the atmosphere. The amount of radiation that is received on the surface of the earth spectral-wise is determined by the constituents of the atmosphere as also the particulate matter that may be intervening. Observations at low angles that is usually the case in terrestrial observation could further aggravate the problem. The weather conditions, such as rain, snow, haze and fog could reduce the clarity of vision. Dust and sand thrown up by vehicular movement and various obscurants, such as smoke could be of special significance in battlefield environment. Presence of pollutants resulting in smog could drastically reduce the visibility down to a few metres. Such conditions do make the contrast rendition quite difficult in the observation plane. It is also well known that astronomical observations are required to be suitably corrected experimentally and theoretically to annul the effect of the intervening atmosphere besides choosing correct locations for observation. The common visible effect observed by the naked eye is the twinkling of the stars. On the surface of the earth atmospheric variation in refractive index may lead to effects like mirage and distortions of varied nature at noon time or in sandy terrain. Yet for quite sometime, the atmosphere does retain a reasonably uniform refractive index value and permits good vision, so much so that the refractive index value is generally assumed to be unity, i.e., same as for vacuum in most calculations. The vision could be excellent on a perfect day as witnessed in Sub-Himalayan terrain which is reasonably free of pollutants and non-atmospheric particulate matter.

An Introduction to Night Vision Technology 2500 SOLAR IRRADIANCE OUTSIDE ATMOSPHERE SOLAR IRRADIANCE AT SEA LEVEL CURVE FOR BLACKBODY AT 5900°K

2000

1500

1000

500 O3 0

VISIBLE SPECTRUM

SPECTRAL IRRADIANCE W m–2 m–1

30

O3 H2O O2, H2O H2O H2O H2O H2O

H2O, CO2 H2O, CO2

H2O, CO2

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0

WAVELENGTHm) Figure 3.1. Spectral radiance of the sun at zenith on earth

The atmospheric effects manifest themselves through absorption, scattering, emission, turbulence and secondary sources of radiation, such as the skylight and reflections from large areas like the clouds and the water masses on the earth. During night, such conditions could be relevant to the moonlight and starlight illumination. The profile of the spectral radiance of the sun at mean earth-sun separation is shown in Fig. 3.1[1]. It also shows the absorption at sea level due to the atmospheric constituents. Obviously transmission is quite significant for the visible region as also for the near infrared, though absorption bands due to water and carbon dioxide do make significant inroads. Likewise Fig. 3.2. shows transmission in percentage value extending right up to 16 m and good transmission in the 3–5 m and 8–14 m bands. These good regions of transmission are also referred to as the atmospheric windows. As we are concerned more with terrestrial, i.e., horizontal transmittance it would be interesting to look at Fig. 3.3, which shows transmittance at sea level containing 5.7 mm precipitable water at 26 °C over an atmospheric path of 1000 ft (@ 305 m)[2,3]. This graphic data also confirms good transmittance in the visible, near infrared 3–5 m, and 8–14 m bands. The data for horizontal transmission would certainly vary significantly dependent on the local condition of observation, but the spectral nature of transmission would by and large be similar.

The Environment ATMOSPHERIC TRANSMISSION

100 TRANSMISSION (%)

80 60 40 20 0

1

2

3

4

6

5

7

8

9

10

11

12

13

14 15

16

WAVELENGTH (m)

Figure 3.2. Atmospheric transmittance vs wavelength

3.2

ATMOSPHERIC ABSORPTION & SCATTERING

The gases that constitute the atmosphere absorb incoming radiation to the planet dependent on their molecular constituents and their characteristics in the spectral bands related to their structure. These gaseous constituents in order of importance are: water vapour, carbon dioxide, ozone, nitrous oxide, carbon monoxide, oxygen, methane and nitrogen. Water vapour and carbon 100 80 60 40

TRANSMITTANCE (%)

20 0

0.5 100

1.0

1.5

2.0

2.5

3.0

3.5

4.0

80 60 40 20 0 100

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5 8.0

8.5 9.0 9.5

80 60 40 20 0 10.0

11.0

12.0

13.0

14.0 15 16 17 18 19 20 21 22 23 24 25 WAVELENGTH (m)

Figure 3.3. Transmission over ~ 305 m (1000 ft) horizontal air path

31

32

An Introduction to Night Vision Technology

dioxide are the most important molecules in this respect while ozone plays a significant part in the absorption of ultra-violet and absorption in the 9-10  m region in the upper layers of the atmosphere. The effect of absorption is attenuation of the signal strength dependent on the wavelength of the incident light. Basically, in absorption, the incident photon is absorbed by an atmospheric molecule, causing a change in the molecule’s internal energy state. The infrared and visible photons may just have adequate energy to enable transition in the rotational or vibrational energy states of a gas molecule. As the energy matching is more for the infrared, absorption is not that significant in the visible region. The absorption due to aerosol would depend on its density. While the energy taken out of a beam of radiation by absorption contributes to the heating of the air, the energy scattered by molecules, aerosol or cloud droplets will be redistributed in the atmosphere. In addition to absorption, the signal strength is further altered due to scattering by air molecules, aerosol and other particulate matter present in the atmosphere. The scattering effects are dependent on the particle size and could be thought of in three categories. The first, where the particle size is relatively small in comparison with the wavelength of the incident light; second, where it is of the same order; and third, where the particle size is relatively large in comparison with the wavelength of the incident light. The relative sizes and their density for important atmospheric constituents are indicated in Table 3.1. Table 3.1. Scattering type Rayleigh Mie/Rayleigh Mie Non-selective

Important atmospheric particles that cause scattering Particle Air molecules Haze particles Fog droplets Rain drops

Radius (m)

Density (per cubic cm)

10 – 4 10 –2 –1 1–10 102–104

1019 10 –103 10–102 10–5–10 – 2

Air molecules as will be observed are relatively much smaller in comparison to the wavelengths of light. The particles of this size would scatter light in all directions thus reducing the signal strength of the incident light but at the same time adding a small amount of forward scatter. As the particle size increases approximately to a size of quarter of a wavelength, the intensity of the scattered light in the forward direction becomes more prominent and much more so when the particles are much larger than the wavelength of light.

The Environment

In such situations, apart from reducing the signal strength, unwanted scattered light in the forward direction is also focused in the focal plane of an instrument system that might be observing an event. The amount of scattered light that would be present would also depend on the parameters of the instrument design, such as its field of view, magnification, entrance aperture size, and its focal length. Though theoretical calculations are not that simple, an approach to the problem is made by defining the attenuation or extinction coefficient as The transmission t = t o e   R

(3.1)

where to is transmission through the vacuum and  is the attenuation coefficient over a path length R in the atmosphere. One could also state that the attenuation or extinction coefficient has an absolute value proportionate to the inverse of the maximum range beyond which there is no transmission of the incident light. This coefficient, which takes into account both losses due to absorption and scattering, is specific to a given wavelength and assumes uniform atmosphere over the pathlength R. In reality, a practical value of this coefficient may be simulated by working out the coefficients for small bands of wavelengths and then appropriately averaging those values over the required spectral range. Experimentally, a number of lasers at different wavelengths could be used for such measurements. This attenuation coefficient would in turn have a contribution from absorption and scattering. Using subscripts a and s respectively for absorption and scattering, we could write down that

  a  s

(3.2)

s in turn, would have contributions from different sizes of particles, each group having particles smaller than the wavelength of the incident light, of the same order as the wavelength of light and of an order where the particle size is relatively much larger. Where the particles are much smaller than the wavelength of light, the problem can be addressed by following Rayleigh’s theory of scattering. In such cases the scattering coefficient can be shown to be proportional to  – 4. The dispersion of scattering about the scattered particle is generally symmetrical showing equal forward and backward scattering. This is true of air molecules. The second

33

34

An Introduction to Night Vision Technology

set, where the particle size is of the same order as the wavelength of the incident light, would primarily be due to aerosol. The scattering in this case becomes a complex function of particle size, shape, refractive index, scattering angle, and wavelength. This could be addressed by utilizing the Mie theory of scattering. In this case, the intensity of the scattered radiation becomes less dependent on wavelength and more dependent on angle, with a distinct peak in the forward direction. In the third group where the particles are much larger than the wavelengths of the incident light, the particles would behave like micro-optical components. Thus their theoretical treatment could be essayed by utilizing the concepts of geometrical optics. This type of scattering also referred to as nonselective scattering or white light scattering (because of lack of dependence of scattering on wavelength) or scattering in the geometrical optics regime could explain scattering due to such large particles as raindrops. Scattering intensity has still a strong angular dependence with a strong peak in the forward direction. 3.2.1

Scattering due to Rain & Snow According to Gilbertson[4], the scattering coefficient in rainfall is independent of wavelength in the visible to far infrared region of the spectrum and could be estimated by the equation

s(rain) = 0.248 t 0.67

(3.3)

where  s(rain) is the scattering coefficient in km , and t = the rainfall rate in mm hr –1. –1

More recent articles by Chimelisx and others[5] give three different formulae for the scattering coefficient due to rain, but all the four formulae are really close enough and do not differ significantly. Empirical relationships have also been developed for snow-based on experimental results and it has been found that the results tend to show two groups, one for snow in small needleshape crystals and the other for larger plate-like crystals. The relationships are as under[6]

s(snow) = 3.2 t

0.91

..for small needle shaped crystals (3.4)

and

s(snow)= 1.3 t 0.5 .. for larger plate like crystals

(3.5)

where the rate of snow accumulation is expressed as equivalent to liquid water rate in mm/hr given by t.

The Environment

3.2.2

Haze & Fog While the relationships developed for scattering coefficient in rain and snow help to a reasonable extent in the estimation of the attenuation or extinction coefficient during such conditions, it may relatively be easier to adopt the concept of visibility which serves as a good substitute for the total attenuation coefficient due to almost all the atmospheric weather variables including haze and fog as the major contributors. Visibility is obviously related to contrast rendition and the contrast sensitivity of the human eye. 3.2.3

Visibility & Contrast The terms visibility, visual range and meteorological range all refer to the horizontal or slant range in which the attenuation of the contrast is not less than 2 per cent, as related to the inherent contrast of an object against its horizon-sky, or its background, i.e., if Co is the inherent contrast of the object, CR the apparent contrast of the object at a distance R, then CR/Co should be 2 per cent or more.

CR /Co  2% defined as

(3.6)

The contrast Co at the object plane, i.e., R = 0 may be

CO 

LO  L BO L BO

(3.7)

where LO is the object luminance, i.e., flux per unit solid angle per unit area or intensity per unit area and LBO is the background luminance. The contrast CR at the observation plane at a distance R could be similarly defined as

CR 

LR  LB LB

(3.8)

where LR is the luminance in the observation plane and LB is the background luminance in the same plane. Equations (3.6) and (3.7) are interrelated as

L R  L o . e  R

(3.9)

L B  L BO . e   R

(3.10)

35

36

An Introduction to Night Vision Technology

The above relationships follow from the fundamental relation of Eqn 3.1 which can also be rewritten as  R =  O.e

– R

,

where  O is the flux radiated by the object and  R is the flux received at distance R, while  is the attenuation coefficient over the same path length. These equations require to be modified to take into account the luminance that is scattered into the line of sight by the rest of the atmosphere. If this scattered luminance is Lin, the modified equations are:

L R  L O . e   R  L in

(3.11)

L B  L BO . e   R  L in

(3.12)

Thus, we have

CR 

(L O  L BO ). e   R LB

and multiplying by

(3.13)

LBO LBO , we have

L   Co  BO  e   R  LB 

using Eqn (3.7)

(3.14)

If the object is viewed against the horizon sky, the background remains more or less the same in both the object plane and the observation plane. The above equation under such conditions reduces to

C R  CO . e   R

(3.15)

For targets against a terrestrial background, the Eqn (3.14) can be remodelled [7] as C R  CO {1  S(1  e σ R )} 1

(3.16)

where S = Lm/LBO, a quantity called sky ground ratio and where Lm is the horizon sky luminance from a direction from which sunlight is scattered along the line of sight and LBO has the same meaning as defined earlier, i.e., background luminance at the object plane. The above equations help in evaluating contrast reduction by the atmosphere and become more practical if related to visibility.

The Environment

Assuming that the resolution of an object is not an impediment in its detection, visibility could approximate to the range at which a black object is just seen against the horizon sky, i.e., at a contrast of 2 per cent. One could therefore rewrite equation (3.15) as

C R / C O  0.02  e   Rv , where Rv is the visibility range. This could also be expressed as

  3.912 / Rv

(3.17)

This relationship is also explained by Fig. 3.4.

ATMOSPHERIC ATTENUATION COEFFICIENT (n) in km –1

5

2

MODERATE FOG

Visibility or visible range is recorded by metrological stations and is a practical guide in many tasks, such as landing and taking off of aircraft, military tactics and logistics and as

LIGHT FOG THIN FOG

1

HAZE

0.5 LIGHT HAZE

0.2 CLEAR

0.1

VERY CLEAR

0.05

RAYLEIGH SCATTERING 310 km

0.02

EXCEPTIONALLY CLEAR

0.01 0.5

1

2

5

10

20

50

100

200

DAYLIGHT VISIBILITY RANGE (RV ) in km

Figure 3.4. Visibility vs atmospheric attenuation coefficient

37

38

An Introduction to Night Vision Technology

a useful parameter in weather prediction. Aerosol models describe a clear and hazy atmosphere corresponding to a visibility of 23 and 5 km, respectively. As the relationship with R is so obvious, R v could be introduced in all the equations involving R. 3.3

ATMOSPHERE MODELLING

The actual attenuation along a path from outside the atmosphere to the point of interest on the surface can be predicted by suitable modelling techniques to a reasonable degree of accuracy using modern computers. Thus, five atmospheres corresponding to tropical, mid-latitude summer, mid-latitude winter, subarctic summer and subarctic winter with two aerosol models corresponding to a visibility of 5 and 23 km for hazy and clear conditions have been suitably worked out. Each of these atmospheres is different in its temperature and pressure variations as also in absorbing gas concentrations. The information is also available in the form of prediction charts. It may further be augmented by giving attenuation coefficients for the laser wavelengths referred to in Table 3.2. Table 3.2. Laser wavelengths used Laser Type

Wavelength (m)

Nitrogen

0.3371

Argon

0.4880

Argon

0.5145

Helium-Neon

0.6328

Ruby

0.6943

Gallium arsenide

0.86

Neodymium in glass

1.06

Erbium in glass

1.536

Helium-Neon Carbon dioxide Water vapour Hydrogen cyanide

3.39 10.591 27.9 337

A well-known model is the Lowtran[8,9] covering a spectral range from 0.25 m to 28 m. One of its later versions has six atmospheres and eight aerosol's allowing 48 combinations for the selection of a user. Models have also been introduced for battlefield obscurants to predict transmission through smoke due

The Environment

to fire, missile smoke and for possible path lengths through smoke and dust [10]. The software could also take into account scattering due to clouds and fog. Scattering in the 8-14 m region is considerably less than that in the visible region of the spectrum; and the principal attenuation mechanism is molecular absorption, particularly that due to water vapour. Thermal detection of an object versus its background may not be affected by the path irradiance as it would essentially be the same, unless one is viewing an airborne object against a sky or a cloudy background. 3.4

INSTRUMENTS, NIGHT VISION & ATMOSPHERICS

The optical systems act as gatherers of radiant or luminous energy and are so designed as to provide maximum possible signal-to-noise ratio and the object-to-background contrast ratio. The image may be presented as a whole as in an image intensifier system or assembled as an electronic display either by image plane scanning or object plane scanning. The object scene may be scanned using a limited number of detectors either sequentially, parallelly, or in both the formats. Staring and focal plane detector-arrays may simplify the process. As the progress in the sensors, detection systems and related electronics has been very significant in recent times, the devices are becoming more and more noise free and more responsive, permitting one to assume that the images are contrast-limited rather than noise-limited in the optical domain, i.e., in visible, near infrared and thermal regions. Atmospheric effects can be better calculated in terms of modulation transfer function (MTF) and assuming linearity, one could estimate the overall MTF of an observation system through the atmosphere. This approach has led to practical results. Thus, in the case of longer focal length systems providing greater angular magnification, atmospheric effects could limit their resolution and blur small details much more than the larger details. Improvements could be possible by spatial frequency filtering and more so by introducing adaptive optics. There is also the possibility that in sensor detection, the weak scattered light may not be recorded due to the limitations of the dynamic range of the sensor. Atmospheric distortions are bound to be magnified. However, magnification beyond a certain point may prove useless, though when atmospheric conditions are weak or moderate an increase in magnification may improve the system

39

40

An Introduction to Night Vision Technology

performance. Turbulence is more significant in the visible region than in the thermal region or during the night. As during the night-time turbulence is generally at a minimum, the maximum attenuation will be due to atmospheric absorption and scattering. The night-time range achieved would depend on the visibility in the image intensifier systems and on the amount of water vapour present in the case of thermal instrumentation, apart from the optical considerations of the instrument design. One might look for better MTF values in the spatial frequencies of interest to the users and involve concepts such as the minimum resolvable temperature difference for improved thermal contrast in the thermal imaging systems. Improvements in spatial frequency and in reduction of noise with better quantum efficiency of photocathodes could give an edge to image intensifier-based systems in conditions of rain and fog.

REFERENCES 1.

Valley, S.L, Handbook of Geophysics and Space Environment. (Airforce Cambridge Research Laboratories, Office of the Aerospace Research, US Air Force. 1965: Also published by McGraw Hill Book Co, New York, 1965).

2.

"Infrared Components Brochure No. 67CM", (Goleta, CA: Santa Barbara Research Center, 1967).

3.

RCA Electro-optics. (Lancaster: RCA Corporation, Technical Series (Section 7), EOH-II, Solid State Division, 1974).

4.

Gilbertson, D.K, Study of Tactical Army Aircraft Landing Systems. Technical Report-ECOM-03367-4, AD-477-727 (Alexandria, Va: Defence Documentation Center. Jan 1966).

5.

Chimelisx, V. "Extinction of CO2 Laser Radiation by Fog & Rain". App. Opt. vol. 21, no. 18, (1982), pp. 3367.

6.

Seagraves, M.A, "Visible and Infrared Transmission through Snow". SPIE Proc. on Atmospheric Effects on Electro-optical, Infrared and Millimeter Wave System Performance. vol. 305, Aug, (1981).

7.

Middleton, W.E.K, Vision through the Atmosphere. (Toronto: University of Toronto Press, 1952).

8.

McClatchey, R.A. et al, Optical Properties of the Atmosphere. AFCRL-71-0279. (Environmental Research Paper. 354 AD726116, May 10, 1971).

The Environment

9.

McClatchey, R.A. et al, Atmospheric Transmittance/Radiance: Computer Code Lowtran 5. AFGL-TR-80-0067. Environmental Research Paper 897. (Air-Force Geophysics Laboratory at Hanscom AFB, Massachusetts. Feb 21, 1980).

10.

Waldman, G. & Wootton, J, Electro-optical Systems Performance Modeling. (Artech House, 1993).

41

CHAPTER 4 NIGHT ILLUMINATION, REFLECTIVITIES & BACKGROUND 4.1

NIGHT ILLUMINATION Though atmospheric parameters may be an impediment in detection and acquisition of an object or a target over significant distances during night, other aspects of environment particularly relative contrasts available at the imaging plane between the object and its background would be more significant in normal night observation. Light in the visible spectrum though never really extinct during nights, certainly varies in intensity due to the presence of the moonlight or starlight under various environmental conditions. Thus, the vision instrumentation for the night can be considered to be quantum starved and appropriate ways and means have to be adopted to make the best use of whatever quanta are available for realisation of an understandable image. The significance of scattered light during night has to be relatively ignored, as its intensity would be far too low. During the day, the scattered light as also the skylight (due to scattering and peaking in blue) though less intense could still illuminate the hollows or obstructions and enable a better depth in seeing shadows or the like. During the night, directional effects may be more prominent. Thus, observation with the moon behind the observer may give a better range and so also observation from a higher level or an aircraft of the ground below under starlight. Illumination under clouds could be a tricky affair. It could sometimes reflect city lights and enable better illumination and at other times it could totally block even the starlight. During a battle, existence of gunfire, flares, and night illuminants used by an army could give a much better chance of viewing the area with night vision devices than could ordinarily be possible under normal conditions. Even the leakage of light from within a closed tank could give away its position from relatively large distances.

44

An Introduction to Night Vision Technology

Atmospheric conditions naturally cause variations in illuminance on the earth’s surface both during day and night. Table 4.1 gives the approximate values of illuminance due to stars, moon and sun when the atmosphere is both clear and cloudy. Table 4.1. Ground-surface illumination Sky condition

Illuminance (1ux)

Starlight-overcast night

10–4

Starlight-clear night

10–3

Quarter moon

10–2

Full moon

10–1

Deep twilight

1

Twilight

10

Dark day

102

Overcast day

103

Full day light

104

Direct sunlight

105

It will be observed that the cloudy conditions cause a reduction in illuminance by an order of a decade or more. The information in Table 4.1 can also be illustrated graphically with reference to sunrise and sunset to show the effect of twilight conditions which penetrate the night illumination marginally for an hour or so before sunrise and after sunset (Fig. 4.1). The values are approximate and illustrative of the likely behaviour of the night conditions. Target and background reflectance is another important parameter. The percentage reflectance values do not change significantly from daytime to night. The earth’s terrain and the surrounding bodies also radiate in the infrared, typically peaking at around 10 m for a mean temperature of 300 °K. One could thus make use of thermal contrasts in a given scene and implement instrumentation using suitable detectors, optics and display techniques. The spatial resolution would be relatively poorer to the visible imagery, but thermal contrast rendering could enable thermal imagery over large distances in normal weather. 4.1.1

Moonlight Moon has almost the same angular size as the sun when observed from the earth. As moonlight is basically reflected sunlight,

Night illumination, reflectivities & background

its spectral characteristics are similar to that of sunlight with relatively very low intensities. The daylight measurements for reflectivity from objects, atmospheric scattering and attenuation and contrast can therefore be adopted as a whole for computational purposes. As a reflector, the light is bound to be partially polarised but its effects on vision systems are not known to be significant, as the instrumentation is not polarisation-preserving. Dependent on weather conditions and on the phases of the moon, the ambient light is much more variable during night than during the day. Elevation changes in the moon’s position add another dimension in the variability of the ground illumination. Full moon at the local midnight has the highest elevation. The illumination changes can also be sudden due to cloud movements. The scene illuminance could change in a matter of minutes. Some of the characteristics of the full moon are summarized in Table 4.2.

  1

FULL MOON CLEAR

ILLUMINANCE (LUX)

10 –1 FULL MOON-PARTIALLY CLOUDY

10 – 2 STARLIT CLEAR STARLIT PARTIALLY CLOUDY 10 –3

STARLIT

STARLIT HEAVY CLOUDS

AVERAGE

10 –4 10 -4 SUNRISE HOURS AFTER TWILIGHT OR BEFORE MORNING LIGHT

SUNSET

0

1

2 2

1

0

Figure 4.1. Expected variation of illuminance from sunset to sunrise under night conditions.

45

An Introduction to Night Vision Technology Table 4.2. Characteristics of full moonlight Characteristic

Value

Albedo (related to surface reflectivity)

0.072

Effective surface temperature

400 °K

Peak self-emitted radiation

7.24 m

Apparent diameter

0.5°

Effective temperature of reflected

5900°K

sunlight (same as that of sunlight)

Figure 4.2 shows radiance from the moon in terms of watts/sq cm/steradian/m over a range from 0.4 m to 2.0 m. It will be observed that the relative energy concentration is much more in the visible region than in the near infrared, as is the case for daylight. The energy content is obviously more than in the case of illumination due to starlight. Another important factor in relation to the intensity due to moonlight is the change in its phase from new moon to full moon

MOONLIGHT RADIANCE WATTS/SQ.CM/STERADIAN/ m

46

10 – 8

10–9 STARLIGHT 10–10

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

WAVELENGTH (m)

Figure 4.2. Night sky radiation vs wavelength

2.0

Night illumination, reflectivities & background

and vice-versa. The reduction factor could be as high as 300. Thus, for the quarter-moon the illuminance value is around one–tenth to that of the full-moon. Figure 4.3 shows a graph indicating the reduction factor of moonlight vis-a-vis its phase changes. It will also be appreciated that as the cloud cover goes on increasing reducing direct moonlight, the relative contribution from the air glow is likely to effectively increase, particularly in the shadows. Thus, the spectrum in the shadows is more akin to that of starlight rather than to sunlight. Scene brightness would be dependant on the incident illumination and its reflectivity.

FULL MOON

NEW MOON

300

REDUCTION FACTOR TO FULL MOON

200

100 80 60 40

20

10 8 6 4

2

1

20

40

60

80

100

120

140

160

MOON PHASE IN DEGREES

Figure 4.3. Reduction factors of moonlight with its phase changes

47

48

An Introduction to Night Vision Technology

4.1.2

Starlight Starlight is really not evenly distributed in the sky because of the concentration of stars in the milky way releasing a good amount of energy in the visible spectrum, as also due to selective spectral distribution of many stars. The intensity at zenith and along the milky way is higher than that elsewhere in the sky. Nonetheless, we can assume the approximate values for ground illuminance in accordance with Table 4.1 as the average illumination. At the same time, it will be observed from Fig. 4.2 that in the starlight the relative radiation content is more in the near infrared than in the visible, i.e., somewhat opposite to that in the case of moonlight. Intensity in each waveband of interest can also be reduced or altered by the type of cloud cover, rain and fog. Scene brightness as in all cases would be dependent on the incident illumination and its reflectivity. Moonless clear night, i.e., a starlit night sky radiance is composed of the following four components within the visual wavelength range: (a) (b) (c) (d)

Extragalactic sources  1 per cent Stars and other galactic sources  22 per cent Zodiacal light  44 per cent Airglow  33 per cent While the contribution due to (a), (b), and (c) result in a spectrum closer to that of a sunlit or a moonlit sky with appropriate alteration in intensity values, the characteristic spectrum of a starlit sky is more due to airglow. In addition to more or less intense lines in some parts of the visible spectrum, the airglow also yields increasing intensity in the near infrared say up to 2.5 m. Thereafter thermal emission of the atmosphere begins to supress it. Hence, a S-20 photocathode[1] with an extended red response or a S-25 has a reasonably good correlation under both moonlight and starlight conditions and is the photocathode of choice in most image intensifier tubes. 4.2

REFLECTIVITY AT NIGHT Reflectance measurements made during day time are equally applicable during night. However, these measurements assume greater significance during night, as it further lowers the amount of low intensity light that is present in the environment. Reflectance by itself during daytime is not significant, as the number of photons reflected is still large enough to be detected by a vision system. No doubt contrast between the object and its background continues to be an important factor at all levels of vision.

Night illumination, reflectivities & background

While extensive measurements have been done on reflectivity from green foliage, grass, leaves (all in various stages of freshness and decay), earth (various types such as yellow, red, brown and loam) and sand both under wet and dry conditions, the data is place- and weather-specific as it differs from place to place, dependent on the climatic and aerosol conditions of the place. The average percentage reflection may be thought of as 20 per cent or so in the visible part of the spectrum and around 50 per cent in the near infrared. For purposes of computation may be, these values can suggest an average situation. It will be observed that the materials to which we have been referring, usually form the background of a scene (Fig. 4.4 and 4.5)[2]. Likewise, reflectivity measurements have also been made on materials that may provide the signal, such as military clothing and paints – green, khaki or the like on vehicles of various types. The percentage reflectivity from military targets as against their backgrounds is generally so designed that either their reflectivity is poor or is matching the background. While the latter approach, i.e., reflectivity of the same order as the background causing merger with the background may be good for stationary targets or for expanses with a similar background such as deserts, poor reflectivity is a better answer where mobility causes changes in the background. Thus, clothing whether woollen or cotton and vehicular paints are usually designed to have a properly selected drab or dull colour so that reflectivity is less than 10 per cent in most of the visible spectra and in the near infrared. One may further evaluate reflectance ratios between a given background, such as rough expanses of land with little greenery, deserts or the like against various items of clothing, paints and pigments. Assuming similar illumination for the target and the background as is the most general situation, the reflectivity ratios can be used to evaluate relative contrast values. This will enable prediction of the type of resolution possible at a given range. Measurements of this nature vis-a-vis their likely backgrounds provides a good input to an instrument designer. These measurements could be on targets and backgrounds specific to a particular territory and to a prospective aggressor. Similar measurement of the background versus targets would also be called for in the thermal region of 8–12 m where the parameters would be the temperature differential and corresponding emissivities. Here, the parametric measurements

49

An Introduction to Night Vision Technology 80 GREEN VEGETATION

70

60 ROUGH CONCRETE

50 REFLECTANCE (%)

50

40 30 20 10

DARK GREEN PAINT

0 –4

–6

–8

1.0

1.2

1.4

1.6

WAVELENGTH (m)

Figure 4.4. Percentage reflections from surfaces of military interest

have to be redefined in terms of minimum resolvable temperature difference for different spatial values as may be defined by an MTF curve for the system. 4.3

THE BACKGROUND A study of natural backgrounds could help detection and also monitoring in displays. The electro-optical imaging is usually displayed on monitors like the cathode ray tubes utilizing phosphors. As the possible range of intensities is quite large (of the order of 10 10), compressive transformations to limit the output to 10 2 or so have been tried both by nature in the human eye and as a result of technological evolution in photography, TV and the like. Thus, from Weber's law analysing the intensity response of the human eye, as observed threshold, intensity change (L) is proportional to intensity (L); we have (L/L) as a constant. This leads to Fechners logarithmatic scale for human vision, i.e., logintensities are on a linear scale with respect to the stimulus in the

Night illumination, reflectivities & background

0.9 0.8 0.7 REFLECTANCE

0.6

SNOW, FRESH SNOW, OLD

VEGETATION

0.5 0.4 LOAM

0.3 0.2 0.1 0

0.4

WATER

0.6

0.8

1.0 2

4

6

8

10

12

14

WAVELENGTH (m)

Figure 4.5. Percentage reflectances of some common surfaces

eye [3]. Though other transformations have also been proposed, the log-scale explanation seems to be acceptable for a reasonable range of intensity variation (Fig. 1.6). The factors that define the image intensity in the background as projected on the retina can be attributed to (a) (b) (c) (d)

Strength and spectral composition of the illuminant Illuminant's position vis-a-vis the scene and the observer The orientation and reflectance of the viewed area, and The reflection function including textural, spectral and specular factors. Measurement and analysis based on the above factors indicates that the distribution of luminances in a natural scene is log-normal, somewhat similar to the nature of visual sensation. By assigning numbers to the possible range of light intensities one can introduce what is called a lightness scale on the log-normal format to encompass proper scene reproduction on a display system. It could also be that detection of man-made objects is therefore relatively easier as these break the monotony of a scene and enhance its response to the visual system. Similar studies for characterizing the background in the 8–12m bands show that for natural terrain the correlation lengths are in the 30-600 m range and that the radiative temperature standard deviation is of the order 1°–7 °K [4]. In this region of the spectrum, it is obvious that reflectance is not of any consequence and that the detection is based on self-emittance.

51

52

An Introduction to Night Vision Technology

4.4

EFFECT ON DESIGN OF VISION DEVICES All the above discussion leads to the fact that in night combat we are likely to feel the need for low light level vision system with an ability to render visible even low contrast between the target and its background. While technology can take care of quite a few such aspects, a priori information is also essential about the nature of targets and backgrounds which can be of significance both to an instruments designer and to an observer for his appropriate training. Training is also essential for an observer to extract the maximum information possible from a monochromatic phosphor display usually peaking in the blue-green regions of the visible spectrum. From the illumination point of view alone, the night vision devices should fully operate at least from full moonlight to cloudy starlight conditions, i.e., over a dynamic range of three decades, without the need for any operator’s adjustment. Further, while operating over this illumination range, it should be able to extract information about low contrast targets also. Design considerations therefore dictate: (a) (b) (c)

Very high intensity amplifications of the very weak optical signals received with as limited intensification of a dark uniform background as possible, i.e., noise limited. High intensity amplification of weak optical signals received without proportionally intensifying its uniform background, and The photocathode used for image intensification purposes should operate right from the visible into the near infrared as far as possible to make full use of light quanta available both under moonlight and starlight conditions apart from its requirement of high quantum efficiency. This makes S–25 a photocathode of choice as already mentioned earlier.

In case of night vision utilizing the 3–5 and 8–12 m bands which are based on detection of the self-emittance of a body, the nature of the daytime or night spectrum is not of much concern even though it is known that the background natural radiance statistics does change according to the presence of hot sources like the sun. Further as the detection is for self-emittance, reflectivity is also not a parameter for observation. These bands are also well-known atmospheric windows, and hence instrument devices in these parts of the spectrum can have a reasonably significant range dependant on the system and its detector characteristics. Quantum detectors are used for detection of the self-emission of targets and their backgrounds in terms of

Night illumination, reflectivities & background

temperature differentials. The system effort is to correlate the thermal contrast rendition with the desired spatial resolution. REFERENCES 1. Hohn, D.U. & Buchtemann, W, "Spectral Radiance in the S-20 Range and Luminance of the Clear and Overcast Sky". Applied Optics, vol. 12, no.1, Jan, (1973), pp. 52-61. 2. Driscell, W. G. (Ed)., Handbook of Optics. Chp. 14. (McGraw Hill Book Company. 1978). 3. Soule, H. V, Electro-optics photography at Low Illumination Levels. (John Wiley & Sons). 4. IRDE Report on creation of test facilities in the 300-ft long hall. (Dehradun, India).

53

CHAPTER 5 OPTICAL CONSIDERATIONS 5.1

INTRODUCTION

While the contribution of a good optical designer is a prime necessity in the successful design of an electro-optical night vision system, the overall system constraints do lay down the basic requirements for an optical system. Understanding of the user requirements on the one hand and the technical possibilities and limits of an optical designer on the other goes a long way in laying down the basis of a successful design and forms one of the main responsibilities of the system designer. These days we do have a library of optical designs. Coupled with the availability of computers and computer software to design and analyse a given or a modified design from such a library, it may be possible to arrive at a desired solution. Alternatively, it is also possible to arrive at a final solution around a preliminary one that the designer might feel workable on the basis of his experience. The analysis could be in terms of optical transfer function, spot diagrams, Strehl definition, Mare'chal criterion, wave front aberration or based on classical geometrical optics approach. The use of appropriate software with compatible computers certainly helps a great deal in arriving at the optimum designs, drastically cutting down the computational time and the time for decision making. As night detection in the broader sense is not just restricted to the visible spectrum only but could also utilize probing into higher wave bands, it may be worth while to refer to the electromagnetic spectrum of interest as given in Table 5.1

56

An Introduction to Night Vision Technology Table 5.1. Detection possibilities Waveband

Wavelength

Frequency

Nature of imaging

400–750 THz

Passive

Visible (Target reflectance from natural sources)

0.75 to 0.4 m

(Target reflectance from laser sources)

Not useful in this region as the position gets easily known

Near infrared (NIR) (Target reflectance from natural sources)

2.0 to 0.75 m

Target reflectance using NIR search – lights

Not in use these days as the NIR searchlights can be detected using appropriate vision systems

Infrared (selfemittance by the targets)

5 to 3 m 14 to 8 m

(Target reflectance from laser sources)

Detection possible

MMW

3 mm

100 GHz

Active

X-band radar

3 cm

10 GHz

Active

UHF TV

10 cm

3 GHz

Tap-able

60 cm

500 MHz

Tap-able

—

—

150–400 THz

—

60–100 THz 21.4–37.5 THz

—

Passive

Active

Passive Passive

Active

Note: Passive imaging does not give away the observer position while active imaging can do so.

Atmospheric absorption rules out the use of wavelengths lower than the visible for vision at a reasonable distance of military interest. The visible and near infrared (NIR) have been linked with detection by means of suitable photocathodes. Photocathodes have

Optical considerations

been developed which respond to the entire visible and NIR regions right up to 1.2 m or so. The technological considerations for imaging using such photocathodes are the same for photons available both in the visible and the NIR. Maximum utilisation of the natural night illumination is thus possible. Dioptric materials like optical glass can also be used though one would have to watch for their absorption characteristics particularly for the NIR. This is not quite so as we shift to detection and image forming in the infrared bands of 3–5 m and 8–12 m, which are also atmospheric windows. Detection is also passive in these windows as it is based on the selfemittance of bodies in the environment, using appropriate quantum detectors for the spectral bands concerned. The detecting area is micron-sized as against quite a few mm or even cm in the case of a photocathode. While an entire image can be focused onto a photocathode, the quantum detectors referred to only see a very small area of the field. In other words scanning techniques have to be introduced to cover the required field of view. Series, parallel, and series-parallel scanning is resorted to as the number of detectors is steadily increased in an x-y format. The more recent development of staring or matrix arrays can dispense with scanning altogether. Thermal energy detectors are also being tried by using matrices of micro-bolometers. Useful dioptric materials in these spectral ranges are: zinc selenide, zinc sulphide, silicon, germanium and the like. Metal mirrors and polygons are also in use for the scanning optics. Appropriate coatings are necessary in all the cases. In still higher wavebands, the techniques are no longer passive and call for illumination of the object and analysis of its reflection. Picturizing an object scene utilizing TV cameras and its subsequent transmission and reception involves a three-fold action, i.e., picturisation, transmission and reception. While the picturisation aspect is dependent on the region of the spectrum used and its corresponding cameras, transmission may utilize UHF band, referred to in Table 5.1. Reception amplifies and modifies the signal received into an appropriate video signal for display on a cathode ray monitor. Picturisation is possible in the visible, visible and near IR, and the higher IR bands in a passive mode during day or night utilizing appropriate objectives and sensors. Thus we have low light level television (LLLTV) systems and Thermal Imaging (TI) systems utilizing appropriate instruments and detectors. Detection ranges can be reasonably large and transmission of such signals may not be opted for. In other words this approach offers an alternate vision system.

57

58

An Introduction to Night Vision Technology

The entire instrumentation which enables vision or display as referred to above is primarily in the domain of optical and electro-optical engineering and is amenable to what may be referred to broadly as optical techniques. 5.2

BASIC REQUIREMENTS All optical systems should be designed to give as perfect an image as possible for a given object scene for identification of individual objects of varying shapes, sizes and colours distinguished from their backgrounds. As all such scenes are an assembly of points, it may be considered that reproduction of these points and their correct juxta-positioning by an optical system would give an overall faithful image. Thus the characteristics of a perfect optical system imaging a given object scene were defined as under: (a) (b)

To every object point there is one and only one image point. To every group of object points in a plane there corresponds a group of image points in a plane. (c) The image is geometrically similar to the object. It was shown from the definitions above that one could trace out the path of rays emanating from each object point through the optical system, following certain fundamental laws. As the rays travel in straight lines, this approach was defined as the subject matter of geometric optics. It also follows from the above and the reversal property of optical rays in a perfect optical system that conjugation is an important property, i.e., the image formed can also be thought of as object which forms an image which is exactly the same object. Further it was soon realised that there are natural limits to the formation of a point image for a point object, as the point image does get blurred to some extent dependent on the aperture of an optical system even when it is a perfect optical system, due to diffraction effects. One could thus think in terms of the progress of a wavefront through an optical system rather than an optical ray which can be thought of as a normal to the wavelets at the points considered. As our knowledge improved about the image formation, one was led to think in terms of point spread function, the line spread function, edge-gradients, the sine wave response, and the square wave response which take diffraction effects also into account. The intention is not only to have a perfect image but also an analysis of the object contrast vis-a-vis its background and as related to resolution. In actual practice, imagery within certain tolerances in relation to a perfect image may be quite acceptable. These tolerances based on practical evaluation of systems may be defined in terms of geometrical or wavefront

Optical considerations

aberrations. Tolerancing could also be in terms of the optical transfer function. Contrast enhancement techniques may additionally be resorted to where object identification from its background presents relative difficulties even in perfect imagery. While the detailed information on these aspects is available in the standard texts on optical design[1], it is our intention herein to restrict to overall assessment and the systems for night vision only. 5.2.1

System Parameters The system parameters that are of overall relevance may be considered to be magnification, focal length, conjugate relations, location of entrance and exit aperture, numerical aperture, vignetting, field of view, tolerancing, consideration for contrast, and resolution. An optical system is generally an assembly of individual lenses, symmetrically normal to a common axis called the optical axis, i.e., the system has rotational symmetry around the optical axis. This assembly is generally well-corrected and forms as perfect an image as is possible. The system may comprise more assemblies involving prisms, erectors, eyepieces, and the like. The rectangular block shown in Fig. 5.1 is the outline of an optical assembly. Conventionally, the light is supposed to fall on an optical system from the left side, also referred to as the object space. The points of our interest are focal points F and F ', the principal points P and P ' and the nodal points N and N'. A parallel ray, i.e., the one parallel to the optical axis from the object FRONT FOCAL PLANE

REAR FRONT PRINCIPAL PRINCIPAL PLANE PLANE

A

Q

A' P'

P



q

F

N



N' B

REAR FOCAL PLANE

q'

F'

B' Q'

U

f

OPTICAL SYSTEM

f' V

Figure 5.1. Imaging sequence through an optical assembly

59

60

An Introduction to Night Vision Technology

space passes through the focus at point F ' which is the focus point in the image space. Likewise, a parallel ray from the image space passes through the focus F in the object space. Focal planes can be defined as planes normal to the optical axis at the focal points. Thus, if the object is assumed to be at infinity or for practical purposes at a reasonably large distance R, then its image would be focused in the focal plane itself. In other words, the conjugate points to all object points at infinity lie in the rear focal plane. If now an object to linear size d0 at infinity subtends an angle  at the optical system, we have in the object space

d0  R

(5.1)

It is of course assumed that the object is at quite a large distance in comparison to the focal length and that the angle  is small enough for tan to be replaced by . As the optical system brings the rays from the object to the focal plane to an image size d i the equivalent or effective focal length gets defined in such a manner that

f

di

(5.2)



Having defined the effective focal length value in these terms, the transverse magnification m of the system can also be defined as

m

di do

(5.3)

Combining Eqns 5.1, 5.2 and 5.3, we have

di 

f . do R

m  f / R

(5.4) (5.5)

These relationships are of interest to a system designer. As already stated, parallel beams from an infinitely distant object are brought to a focus in the focal plane. While doing so, the beams undergo deviation at each and every optical surface of the assembly and then emerge from the last surface to come on to the focal plane. Principal planes or surfaces are defined as the unique imaginary surfaces from which these parallel beams could

Optical considerations

have been singly refracted to come to the same focus. There are two such surfaces in each assembly depending on whether the parallel beam is incident from the object space (A’P’B’) or the image space (APB). Their intercepts on the optical axis are the principal points P and P'. These surfaces and points are indicated in Fig. 5.1. The effective focal length (EFL) is defined as the distance P'F' and PF. It will be observed that this definition tallies with the definition as per Eqn 5.2 for P'F' = f '. Likewise, the nodal surfaces and points are defined as the two imaginary surfaces and their intercepts on the optical axis wherein if a ray is incident from an object point, the same is refracted without any deviation from the corresponding nodal point of the second nodal surface. Thus, in Fig. 5.1, the ray QN is transmitted parallel to itself as N'Q'. The focal, principal and nodal points are referred to as cardinal points of an optical assembly or subsystem. Back focal-length and front focal-length are measured in terms of the distances from the rear and front surfaces to their respective focal points. These measurements are important while going in for the mechanical design of the subsystem. Other important parameters for correct placements are the edge and centre thicknesses of all the optical elements and their interdistances. We may now proceed to define the field of view (FOV). The FOV refers to the angle over which ray bundles are accepted from the object space by the lens system. This angle is restricted by the field stop in an image plane which for distant objects is just the back focal FRONT PRINCIPAL PLANE ENTRANCE PUPIL

REAR PRINCIPAL PLANE

FIELD STOP IN BACK FOCAL PLANE

FOV di

Figure 5.2. Field of view in relation to field stop

61

62

An Introduction to Night Vision Technology

plane (Fig. 5.2). The field stop can be placed in any real image plane in a relaying system of optics, to give a sharp boundary to the FOV. It will be observed that (5.6)

tan ( FOV / 2 )  d i / 2 f

where di is the linear dimension of the circular field stop. The field stop can be rectangular also in which case the FOV will have two different values in the corresponding perpendicular directions. An aperture stop may also be introduced into an optical system to physically limit the size of a parallel bundle that enters it. Usually the aperture stop is the clear aperture of the front surface but it can be anywhere based on the design consideration. The image of the aperture stop in all the system elements preceding it is called the entrance pupil and in succeeding elements it is the exit pupil. Both the aperture and field stops are of importance as one limits the size of the parallel beam bundle and the other the angle of entry of such bundles. Parallel bundle's size determines the brightness in the image while entry at greater angles requires a much more stricter control of aberrations. Further, at greater angles of incidence, the entire beam may not find an entry to the image plane, as a mismatch between the entrance pupil and the field stop may limit its transmission through all the optical elements of the system. This is referred to as vignetting, and leads to a greater loss of brightness towards the edges of the image field. Vignetting becomes a serious problem in night vision systems. Relative aperture or F number is also a relevant parameter from the system point of view. It is defined by f '/D where D is diameter of the entrance aperture (Fig. 5.3).

ENTRANCE APERTURE

PRINCIPAL SURFACE A'

f' D

P'



B' PARALLEL BEAM FROM INFINITY

OPTICAL SYSTEM

Figure 5.3. F number and numerical aperture

F'

Optical considerations

F number  f '/D

(5.7)

Similarly, numerical aperture (NA) by definition is the sine of the angle that the marginal ray makes with the optical axis, i.e., for angle/2 in Fig. 5.3. As the principal surface of a perfect optical system is defined as the imaginary single surface from which after refraction a parallel beam comes to a focus in the focal plane, i.e., the image plane for a distant object, it is obvious that this would be a segment of a sphere centred on the image point. Thus, we have D 2f '

(5.8)

1 2NA

(5.9)

NA  sin /2 

and

F number 

Both these values are of importance in objective systems, as these decide the light gathering power of the system or its throughput. In systems design, matching of the throughputs may be quite essential particularly, where it is the intention to collect as much light as possible and then be able to transfer it to the next assembly in the chain without any loss. Obviously, at unit magnification, all the subsystems should have the same numerical aperture. Nevertheless, practical demands will have to be met where some magnification is also desired. Referring to Fig. 5.4, it will be observed that the numerical aperture in the object space is

sin  / 2 

D 2

D 2u

U

(5.10)

V



'

D

OPTICAL SYSTEM

Figure 5.4. Matching of numerical aperture

63

64

An Introduction to Night Vision Technology

and in the image space is

sin   / 2 

D 2v

(5.11)

As the principal surfaces are segments of spheres centered on object and image points on the axis respectively, we thus have v sin  / 2   m (magnification ) u sin   / 2

(5.12)

This means that it is possible to determine limiting values for throughput and selecting reasonable values for magnification and conjugate distances before one goes into the detailed optical design. 5.2.2.

Design Approach

Snell’s Law n1 sin1 = n2 sin2 enables one to find out the direction of a light ray after refraction at the interface between two homogenous, isotropic media of differing indices of refraction where n1 and n2 refer to the refractive indices of the media before and after refraction and likewise 1 and 2 refer to angle of incidence in the first medium and angle of refraction in the second. The sine function in the Snell’s Law can be expanded into an infinite series based on the formula: Sin  =  –  3/3! +  5/5! –  7/7! +  9/9!......

(5.13)

If the sine function is replaced by  and the refraction of rays worked out from the object to the image plane on this basis through an optical system, it is called the first order or paraxial approach. Obviously, it is paraxial as it is only close to axis that sin  can be approximated to . Formulae have been developed for the purpose and used in the first stages of design of an optical system to determine parameters like the system focal length, magnification, conjugate distances, etc. The next step in closer approximation includes the third order term in the sine expansion. As it was mainly investigated by Seidel, the aberrations resulting from this approach are referred to as Seidel or third order aberrations. Restricting to monochromatic light, these aberrations have been classified as spherical aberration, astigmatism, field curvature, coma and distortion. Formulae and tolerances have been worked out and methods developed to

Optical considerations

annul these aberrations as far as possible. It is possible to overcome chromatic aberrations, i.e., aberrations due to various colours in the white light because the index of refraction of a material is a function of wavelength thus offering a possibility to balance chromatic differences by appropriate selection of refractive index values for each lens or a prism in an optical system. It is not quite easy mathematically to involve the higher terms, i.e., the fifth order and onwards for better correction and hence where the user demand is more stringent it is essential to go in for exact trigonometrical ray tracing. This would be particularly true for the systems involving large angular FOV and large apertures demanding a state of very high correction. Even if the geometrical approach as discussed above were to result in an exact point image for a point object, the diffraction of light through the optical system results in a spread of the point image dependent on the diameter D of the entrance aperture, the focal length f of the system and on the wavelength  of the light used. The point image spreads into an intense circular patch surrounded by alternating dark and bright rings. The maximum energy, i.e., 83.9 per cent is concentrated in the central circular patch, 7.1 per cent in the first bright ring and the rest 9 per cent spread over the remaining rings in declining order (Fig. 5.5). For 7.1 % MORE IN FIRST BRIGHT RING

NORMALIZED PATTERN IRRADIANCE

83.9% IN AIRY DISC 1.0 .9 .8 .7 .6 .5 .4 .3

CIRCULAR APERTURE

.2 .1 0.0

–8 –7

-6

–5 –4

–3 –2

–1

0

1

2

3

POSITION IN IMAGE PLANE

4

5

6

7

8

Figure 5.5. Diffraction pattern of a point object through a circular lens system.

65

66

An Introduction to Night Vision Technology

practical purposes therefore the size of image point is equal to the diameter of the central bright spot which is referred to as Airy’s disc and whose radius is given by

1.22  .f (5.14) D where r is the radius of the Airy disc, the practical size of an image point. r 

In most of the cases, the systems are aberration limited, i.e., though the aberrations are within specified tolerances, the total effect of these tolerances is over and above the diffraction effects. However, in quite a few cases it does become necessary to minimise all the aberrations so that the total wavefront aberration is of the order of the practical limits as may be laid down by diffraction effects. Of course further confinement of the Airy’s disc beyond the diffraction limit is not possible for a given diameter, though techniques do exist for detecting weak signals in the neighbourhood of a very strong signal. 5.2.3

Design Evaluation Optical components, subsystems, and systems can be evaluated by a large number of techniques for their parameters and aberration characteristics involving precision opto-mechanical instrumentation, collimators, auto-collimators, interferometers, modulation transfer function (MTF) measuring equipment, and the like. While discussion of all these techniques is beyond the scope of this book, reference to the MTF approach is particularly significant from the systems point of view for night vision devices. An incoherent imaging system can be characterized by a two dimensional optical transfer function (OTF). The OTF is a complex quantity whose modulus is a sine-wave amplitude response function called the MTF and whose argument is the phase transfer function (PTF). Thus







 



OTF Vx ,Vy  MTF Vx ,Vy exp jPTF VxVy



(5.15)

where Vx and Vy refer to spatial frequencies in the two imaging directions of the image of an isoplanatic patch. The MTF gives the modulation reduction of the imaging system versus its spatial frequency when a sinusoidal radiance pattern is imaged. For a perfect imaging system, the modulation transfer function would be unity at all the spatial frequencies of a sinusoidal radiation pattern. However, as we will see, it cannot be so even for a perfect diffraction limited optical system.

Optical considerations

As by and large optical and electro-optical systems are known to behave linearly, it can be shown that the total performance of a complete optical system or an electro-optical system composed of many sub-assemblies is obtained by multiplying the individual OTFs of each of the sub-assemblies. Thus, the OTF of a complete night vision system could be a multiplication of its values for the objective, image intensifier tube and the eye piece or a viewing system. Generally this could be true of MTF values also. The sinusoidal chart as an object instead of a bar chart is certainly more deterministic of the optical system, as the results from it combine the characteristics of contrast and resolution which are otherwise evaluated separately. Illuminating a sine chart uniformly and using it as an object, one can define the object contrast or modulation for the frequency u of the sine chart as

Co 

O max  O min O max  O min

(5.16)

where Co is the object contrast, Omax the maximum transmission through the sine wave chart and Omin the minimum transmission through the same chart for a frequency u. Likewise, the image contrast Ci of the imaged sine wave chart can be defined as

Ci 

I max  I min I max  I min

(5.17)

where Imax and Imin are the maximum and minimum irradiances. The MTF at frequency u is then defined as

MTF (u)  Ci / Co

(5.18)

The MTF curve is arrived at by plotting MTF values against frequency. Instrumentation is available for generating the sine wave objects or simulating their output as also to plot the MTF curves from intensity measurements in the image plane. Instruments are also designed to measure the polychromatic MTF directly. A perfect optical system theoretically would have MTF value of unity at all frequencies, as both Co and Ci would have a value of unity. In practice, however, even a diffraction limited optical system cannot have a unit value for all the frequencies as the aperture of the imaging systems leads to diffraction effects. For instance, it can be shown that for a circular aperture (most usual with lens systems) the monochromatic diffraction limited MTF is given by

67

An Introduction to Night Vision Technology

MTF diffraction limited 

cos

2



1

n  n 1 n 2



(5.19)

where n is the normalised spatial frequency, i.e., the ratio of the absolute spatial frequency u to the cutoff frequency uc due to diffraction, i.e.,

n

u uc

(5.20)

The cutoff frequency is that frequency at which the MTF value is zero. Frequencies may be expressed in cycles per mm (c/mm) or cycles per milliradian (c/mr), keeping due regard to the units used for other parameters. There are several formulas for uc. The one relating directly to Airy’s disc is given by u c (c /mm ) 

D 1.22  f r

(5.21)

1.0

0. 8

IDEAL MTF 0. 6 MTF

68

0. 4

0. 2

0

LENS WITH 1/4 WAVELENGTH ABERRATION

0.2

0.4 0.6 NORMALIZED SPATIAL FREQUENCY

0.8

Figure 5.6. Diffraction limited MTF

1.0

Optical considerations

where D is the diameter of the entrance pupil, f the focal length,  the wavelength of the light, and r the radius of the Airy’s disc, all in mm. The formula can also put into angular terms when

u c (c /mr ) 

D





1.22

r

(5.22)

Where r refers to half the angle subtended by Airy’s disc at the entrance pupil diameter D. Figure 5.6 shows the MTF values plotted against the normalized spatial frequency, n, it will be observed that the diffraction limited ideal performance curve is almost a straight line which dips slightly towards the origin. Comparison has also been made with a lens system that has a quarter wavelength aberration[2]. Obviously all real lenses will have their graphs between the origin and the line indicating the diffraction limited ideal performance. As indicated earlier, one could now evaluate the MTF curves for the objective and the image intensifier tube to arrive at the combined MTF of the system. Nonetheless, experiments following Johnson’s criterion seem to prefer square-wave spatial frequency amplitude response, i.e., bar chart in practice defined by line-pairs per mm. As discussed earlier in Chapter 2, this permits a correlation between acquisition, recognition and detection though these values cannot be cascaded in the manner that is possible with MTF values, in respect of optics, image intensifier tubes, camera tubes, video amplifiers and displays. Some workers have developed calculating and graphical schemes to convert one set of values to the other. The manufacturers of image intensifier tube give the data generally in terms of resolution in line-pairs per mm as also normalised MTF values in line-pairs per mm. 5.3

OPTICAL CONSIDERATIONS It may be better to think of night vision systems as an assembly of an objective, image intensifier tube and an eyepiece or a display system, i.e., it may not be thought of in the conventional manner as a telescopic system though it does behave like one and essentially does the same task. The objective in this case is to collect as many photons as possible from the night sky and concentrate these photons in as small an area as possible so that the intensity per unit area is as high as possible to enable maximum excitation of the photocathode of the image intensifier tube. At the same time one has to reconcile these requirements, with the FOV and overall magnification that may be

69

70

An Introduction to Night Vision Technology

desired from the entire system to design a reasonably aberrationfree optical system. In other words, an objective has to be designed both with a high aperture and fast lens system optimised for a given FOV and required overall magnification usually to go with standard unit magnification I.I. tubes of 25 mm and 18 mm diameters, in most cases. Thus the objective designs are close to fast photographic objectives and do make use of knowledge available in that field. The objectives could be totally refractive, catadioptric, or catoptic, i.e., purely lens system, mixture of lenses and reflecting surfaces or only based on reflecting surfaces. Refractive lenses have been characterised and generally developed from the well known basic configurations of Petzval lenses, the Triplet, and the Symmetrical. Petzval lenses took the form of two doublets to form a system, though in some cases each of the doublet was replaced by a doublet with associated singlets. The main advantage of this lens system is that it lends itself to comparatively high apertures though with a considerably reduced field coverage. It may be preferred where field flattners can also be introduced and the desire is to have a relatively flatter field. It has been mainly used for cine-film projection. Triplet lens family started with the Cooke's Triplet and in its earlier form comprised a single negative lens placed between two positive lenses. Lenses of the triplet and derived forms are used mainly for narrow to moderate field of view. It can be stated that higher the aperture of the triplet, the smaller is the field of view that it covers. For better aberration control and balancing, element splitting and compounding of three elements of a basic triplet have led to a number of useful photographic objectives like the Tessar, Pentac, Sonnars, and the like. Symmetrical systems as the name implies are symmetric with respect to a central stop where the front and rear sections are related to each other as mirror images. Apparently, such a system results in balancing out of a number of aberrations particularly at equi-conjugate positions. The system is altered somewhat in its symmetry so that it can usefully operate at infinite conjugates also. Complex systems, such as Dagor, Aviogon, Topogon and the like have resulted from it. Double Gauss system (six piece lens system) developed in this category have resulted in high aperture lenses of moderate field of view, i.e., of the order of 60° or so and this has made them adaptable to a number of applications. With advancements in lens design, the distinction between the basic triplet and the symmetrical lenses has tended to become blurred; nonetheless, the approach is still significant and valuable.

Optical considerations

Utilization of a reflecting curved surface as the primary focusing element in an optical system leads to the development of catoptic and catadioptric systems. To correct spherical aberration of a concave mirror, some designs utilize a parabolic surface but then the useful field may be somewhat restricted. Thus, a paraboloid primary and a hyperbolic secondary, a catoptic system referred to as the Cassegarainian system has resulted in a successful objective and led to better systems for special applications. Likewise catadioptric systems involving a primary concave mirror with suitable correctors in front and field flattners has also resulted in systems of interest. In Mangin mirrors, the back surface of a double concave lens is used as the primary mirror permitting its use as a refracting component also to enable better correction at higher apertures. Concentric designs were also introduced by Bouwers[3] for realisation of useful high numerical aperture systems. Aspheric correctors – the Schmidts, have also been used in appropriate planes in front of the primary concave mirrors to replace the set of spherical correctors that are otherwise needed. The problem of accessibility to the focal plane and the likely long length of systems utilizing a primary concave mirror are variously solved by different designers. From the MTF point of view, lens systems may be corrected to achieve different specifications for different applications. Thus, for photographic objectives, these may be designed for optimum performance at higher spatial frequencies. That however would not be the case for the night systems, i.e., lens system coupled to an image intensifier tube or to a low light level imaging TV as the highest possible MTF values will be desired in the lower range of spatial frequencies because of the limitations imposed by photoelectron statistics and the nature of fibre-optics elements that have been used. Figures 5.7a, b and c shows some illustrative optical designs developed and actually utilized in instrument systems at the Instrument Research & Development Establishment at Dehradun, India. One can easily decipher the triplet and its derivative, the symmetrical system, and the catadioptric systems including one utilizing a Mangin mirror and the other a Schmidt corrector. It also shows the square-wave frequency response of each system at different field angles. Classical aberrations can also be appreciated by visualizing the size of the spot diagrams. It can be observed that the progress is towards faster F numbers at larger apertures[4].

71

80

160

70

100

80

INTERMEDIATE

40

40

20

AXIAL 80

100

0

20

40

MTF 60

80

100

0

20

MTF 60 40

80

100

20

2.8°

0°

40 60 LINE PAIR/mm

0°

40 60 LINE PAIR/mm

7°

0°

40 60 80 LINE PAIR/mm

20

10°

20

16°

FREQUENCY RESPONSE

0 22°

20

40

MTF 60

Figure 5.7a. Some illustrative optical designs

F/1.7 INFRARED ARTY SIGHT OBJECTIVE

F/2 TV OBJECTIVE DOUBLE GAUSS

F/6.3 PHOTOGRAPHIC TRIPLET

80

FULL

SPOT DIAGRAMS

80

80

72 An Introduction to Night Vision Technology

25

100

15

15

0

20

MTF 60 40

80

100

100 80 MTF 60 40 20 0

Figure 5.7b. Some illustrative optical designs

F/1.2 PASSIVE NIGHT PERISCOPE OBJECTIVE (50 mm DIOPTRIC)

150

F/1.3 PASSIVE NIGHT TELESCOPE OBJECTIVE (80 mm CATADIOPTRIC)

30

25° 10 20 30 LINE PAIR/mm

17°

0°

10 20 LINE PAIR/mm

30

Optical considerations 73

130

1

12

3.5

10

100 MTF 80 60 40 20 0

100 80 MTF 60 40 20 0

40

Figure 5.7c. Some illustrative optical designs

F/1 OBJECTIVE WITH ASPHERIC CORRECTOR (200 mm CATADIOPTRIC)

F/1 PASSIVE NIGHT TELESCOPE OBJECTIVE 200 mm

200

0° 2.8°

10 20 LINE PAIR/mm

10 20 LINE PAIR/mm

3.5°

30

30

0° 2.5° 8°

74 An Introduction to Night Vision Technology

Optical considerations

REFERENCES 1. Cox, Arthur. A System of Optical Design. (The Focal Press, 1964). 2. Griot, Mells. Optics Guide 5. 3. Bouwers, A. Achievements in Optics. (New York: Elsevier Publishing Company Inc., 1950). 4. Various Optical Designs and their Characteristics. (IRDE, Dehradun).

75

CHAPTER 6 PHOTOEMISSION 6.1

INTRODUCTION The need for detection of weak radiation signals both in visible and the infrared has, of necessity, led to the development of quantum detectors. Quantum detection may be based on the principles of photoemission or utilize solid-state devices in which the excited charge is transported within the solid either as electrons or as holes. Photoemission of electrons has been utilized in image intensifiers (I.I. tubes), photomultipliers and the like, or in general, in various vacuum or gas-filled tube devices for different applications. Solid-state devices may be classified as photoconductive or photovoltaic. These may be simple p-n junctions, photocells, phototransistors, avalanche photodiodes, p-i-n photodetectors, schottky-barriers, or quantum well devices. Photoemissive surfaces are possible in relatively larger sensitive sizes. 6.2

PHOTOEMISSION & ITS THEORETICAL CONSIDERATIONS Materials (metals, metal compounds or semiconductors) which give a measurable number of photoelectrons when light is incident on them, form photocathodes in a vacuum tube enveloping both cathode and anode in an electric circuit (Fig. 6.1). The electrons emitted from the photocathode when the light is incident are collected at the anode maintaining the flow of the current as the anode is positively charged. As the anode potential is increased, the current also increases which ultimately reaches a saturation value beyond which further increase of the anode potential is not helpful. This saturation value of the current is proportional to the intensity of the light incident on the photocathode. If the anode potential is now reduced, the current value can be reduced to zero at a negative threshold potential. This potential value is found to be dependent on the wavelength of the incident radiation and not on its intensity.

78

An Introduction to Night Vision Technology LIGHT VACUUM ENVELOPE (GLASS) PHOTOCATHODE

ANODE

BATTERY

GALVANOMETER

Figure 6.1. A basic photoelectric circuit

6.2.1

Theoretical Consideration The energy of the incident photon Eph is given by E ph  h ν

(6.1)

where h is the Planck’s constant given by h = 6.624  10 Js and ν is the frequency of the incident light. Assuming that the electron in the material has a maximum kinetic energy, W 1 and it has to spend an energy W for its release from the material by overcoming the potential barrier at the cathode surface, then according to the quantum approach the maximum energy that the photoelectron can possess is given by –34

E  hν  W1  W

(6.2)

If  is the work function, i.e., the value of the potential barrier measured in volts at the cathode surface, we have according to the thermionic theory e  = W  W1

(6.3)

Where e = 1.59  10 –19 Coloumbs is the charge of the electron. Eqn 6.2 can now be rewritten as E 

hc  e

(6.4)

Photoemission

where E is the maximum emission velocity of the photoelectron measured in electron-volts and  is the wavelength measured in m, substituting the value h, e and c (the velocity of light = 2.99 108 m/s) we have E =

1.246 10 6



 

(6.5)

From Eqn 6.5, we can easily deduce the maximum value of 0 as

o =

1.246 10 6



(in m)

(6.6)

The lower the value of the work function, farther the wavelength threshold is shifted towards the longer wavelengths. For a given value of , the work function, the maximum value of wavelength  also gets fixed above which photoelectrons do not acquire any energy to escape from the photocathode. This quantum approach due to Einstein explained these experimental observations. 6.2.2

Types of Photocathodes & their Efficiencies

Efficiency of photocathodes is expressed in terms of their quantum yield. If each incident photon were to generate one photoelectron, the quantum yield is said to be unity or 100 per cent. In practice the yields are much lower. Apart from photon losses due to reflection of photons from the photocathode surface, the emission of photoelectrons would depend on the optical absorption coefficient, electron scattering mechanisms, and the potential barrier at the surface that has to be overcome. These parameters have been investigated in depth by Spicer and Go'mez[1]. The results obtained show a better understanding of photoemission for both fundamental and practical applications. Figure 6.2 based on their results is very illustrative. It expresses quantum yield for metals, semiconductors, transferred electron cathodes and negative affinity photocathodes against theoretical estimates for response time. In the case of a transferred electron cathode it may be possible to achieve both higher quantum efficiency and faster response. Metals as a group have the electronelectron as the dominating mode of scattering in which case the escape depth for the electron is short. Thus the quantum yield is poor and the time response is very fast. In a semiconductor with a sufficiently low electron affinity there is no electron-electron scattering near the threshold. In semiconductors, a finite band gap separates the highest states filled with large number of electrons

79

An Introduction to Night Vision Technology 10–7

REPRESENTATIVE GaAS (Cs.0)

10–8 THEORETICAL ESTIMATE OF RESPONSE TIME (S)

80

NEGATIVE AFFINITY

10–9 10–10

TRANSFERED ELECTRON CATHODE

10–11

(INCLUDING COMPOSITES, ALLOYS MULTIALKALIS AND ANTIMONIDES)

10–12 10–13 10–14 METALS

10–15

(INCLUDING PURE ALKALIS)

10–16 10–17 10–5

10–4

10–3

10–2

10–1

100

YIELD (ELECTRONS/PHOTONS)

Figure 6.2. Types of photocathodes yield vs response time

and the lowest conduction band states, so that electrons must have sufficient energy above the conduction band minimum to suffer electron-electron scattering. The dominating mode of scattering is the electron-photon (lattice) scattering. Thus it has relatively a larger escape depth for the photoelectrons. The quantum yield is better and the response time is relatively slower in comparison to that of metals. The next type of photocathodes to emerge were the negative electron affinity (NEA) photocathodes. In these photocathodes, the vacuum level is dropped below the conduction band minimum so that electron affinity takes on a negative value. The large response near the threshold is due to the fact that electrons which are inelastically scattered may escape even if they thermalize into the bottom of the conduction band, i.e., in addition to a fraction of photoelectrons escaping without losing their initial energy; most of the electrons thermalize, diffuse to the surface, and escape without losing all their initial energy. By combining negative affinity approaches with structures that allow an internal potential to be applied across the semiconductor nearest the surface, it is possible to extend farther into the infrared (1.4 m). These cathodes have been referred to as transferred electron (fieldassisted) photocathodes. The response time is also faster than the NEA photocathodes. Further researches may lead to better yield, still faster responses and further extension of the wavelength beyond 1.4 m (Fig. 6.2).

Photoemission

6.3

DEVELOPMENT OF PHOTOCATHODES Historically Hertz in 1887 was the first to note the photoelectric effect ahead of the discovery of the electron. Its explanation based on the quantum theory was due to Einstein in 1905. Nevertheless the phenomenon was not fully explained till a much later date when it was realised that photoelectrons are not just surface-emitted but can come from the depth of the material also. The later theories took into account the mechanism of optical absorption of the incident light quanta, transport to the surface from the depth of the material and its subsequent escape from the surface. In earlier days, i.e., up to 1930s the quantum yield or efficiency of the materials (metals) used was less than 0.01 per cent and hence these materials did not prove to be of any practical use. 6.3.1

Composite Photocathodes It was just a little later that Koller and Campbell[2] accidentally and independently discovered that a combination of silver, oxygen and caesium (Ag-O-Cs, also called S-1) produced a photocathode with a better quantum efficiency than known hitherto. The peak quantum efficiency was around 0.5 per cent, at least a decade better than the earlier compounds. It is made by oxidizing a clean silver layer, and distilling caesium vapour into the tube at a temperature of approximately 150 °C. The vapour then reacts with the silver oxide, resulting in a layer of caesium atoms adsorbed on caesium oxide on silver. Evaporation of a further thin layer of silver and subsequent baking causes further increase in sensitivity. The process known as activation with caesium or cessiation in one form or the other continues to the present day for all photocathodes. Because of low work function of these materials, the response extends into the infrared up to 1.2 m. This S-1 photocathode found many applications in industry, astronomy, biology, etc. The layers could be made sufficiently thin to be semi-transparent for special uses in image converters or television camera tubes. This way light can fall on the photocathode from the back while the photoelectrons are liberated from the front side inside the vacuum. This family has also been referred to as the composite cathodes. The image converters are of particular interest to us, as these photocathodes were utilized in such tubes with success to result in images for hot objects and for night vision before the image intensifiers appeared in the military market. 6.3.2

Alloy Photocathodes The next important photocathode developed was antimony-caesium cathode by Gorlich in 1936, a semiconducting

81

82

An Introduction to Night Vision Technology

alloy of formula Cs3Sb. This and similar compound photocathodes are known as alloy photocathodes. It consists of a layer of antimony onto which caesium is distilled at a temperature around 150 °C to form the alloy. Final oxidation resulted in further increase on sensitivity. This cathode is also known as S-11. The peak quantum efficiency was around 15 per cent. It found its use in TV camera tubes, photomultipliers and the like. Another important alloy photocathode was soon developed by Sommers in 1938. It consisted of a base layer of a 50 per cent alloy of silver and bismuth, oxidized and treated with caesium. It resulted in a panchromatic cathode without the high peak of S-11 in the blue and the infrared response of S-1. Though its peak response was less than that of S-11, its relative panchromacity in the visible was of greater use for colour television. The cathode also known as S-10 (Bi-Ag-O-Cs) was used in image orthicons, which were later replaced by photoconductive tubes of the vidicon type. 6.3.3

Alkali Photocathodes

A layer of an alkali metal deposited in vacuum by evaporation or by electrolysis onto a glass envelope which forms the base of a vacuum tube, results in an alkali photocathode. These atoms have a low ionization potential and get absorbed on the base surface reducing their work function and permitting a better photoelectric effect at certain wavelengths. The quantum efficiency at the peak value could be around 10 per cent. Next, multi-alkali photocathodes were introduced by Sommers. A base layer of antimony is first treated with potassium. Post baking, it is mostly replaced by sodium which in turn is partly replaced by caesium. The photocathode of chemical composition Na2KSb (Cs) for the first time showed a reasonably high quantum efficiency of 20 per cent at the peak wavelength. It came to be known as S-20. Its activation process is more complicated but it proved its utility in earlier image intensifier tubes for night vision. Many other photocathodes of the type Na2CsSb, K2CsSb also got developed. K2CsSb after superficial oxidation was almost similar in its response to Cs3Sb (S-11) and proved better for scintillation counters, while Na2K Sb (not cessiated) had a response identical to S-11. The extension of the threshold of the red response from 800 to 900 nm is possible by increasing the thickness of the multialkali photocathodes and/or using different processing techniques leading to an extended red multi-alkali (ERMA) or S-25 photocathode.

Photoemission

6.3.4

Negative Affinity Photocathodes Around this time, i.e., in late fifties a better theoretical understanding of photo-emission as a bulk phenomenon revealed quite a few facts. For instance, in metals the quantum efficiencies are very low due to strong interactions between the photo excited and conduction electrons which limit the diffusion length. In multi-alkalis, photocathodes showed that the yield of S-25 over S-20 is primarily due to a reduction in the effective surface work function. It was also surmised that by the use of caesium it is possible to decrease the surface vacuum level thus improving the yield by orders of magnitude. The electron affinity values could be made lower and lower. Later, it was found that surface treatment with Cs plus oxygen gave an even lower electron affinity and higher average quantum yields, over the visible and the near infrared portions of the spectrum. It became clear that an optimum photocathode may first be realised by selecting a material with compatible optical properties of absorption as also electronic properties that assist photoelectrons ejection (diffusion to the conduction band minimum) and then processing the material surface to reduce the effective electron affinity to a minimum. Table 6.1 shows the advantage of decreasing value of electron affinity on the increase in quantum yield. All the materials in this list have bulk p-type conductivities. The negative value of electron affinity will depend on the height of the band bending (Fig. 6.3). The closer the vacuum level to the valence band, the more will be the bending and consequently greater the reduction in the value of electron affinity. Table 6.1. Photocathode types in relation to electron affinity Photocathode material

Type

Cs3Sb

S-11

Bi-Ag-O-Cs

S-10

Na2KSb (cessiated)

S-20

Electron affinity (eV)

Max. quantum yield in the visible

1.6

0.45 0.9

1.0

0.55

~20 ~10 ~25

650

0.7

Multi-alkali S-25 antimonides ERMA (cessiated)

1.1

0.24

>25

950

GaAs (Cs,O)

1.4

–ve value >35 with respect to bulk conduction band minimum

~1000

NEA

Band gap energy (eV)

Wavelength cutoff (nm) 750 850

83

84

An Introduction to Night Vision Technology

Thus a new era of engineered photocathode materials started, as against the Edisonian research which was prevalent thus far. The negative electron affinity photocathodes were the first scientifically engineered materials of bulk p-type with n-type surfaces where the band-bending could be downwards introducing negative electron affinity. These considerations led to the suggestion that optimally heavily doped (say Zn-doped) single GaAs with a surface film of caesium and oxygen should have better sensitivity than all previous materials. The same was experimentally verified. High doping (~1019/cm3) is used to minimise the band bending region, while the addition of a caesium monolayer allows the vacuum level to drop below the conduction band minimum (Fig. 6.3). Most NEA photocathodes these days are activated with Cs and oxygen, forming a monolayer of caesium oxide and are referred to as Generation-3 photocathodes. For many devices semi-transparent photo-emissive materials are required as in the case of image intensifiers. This requirement is met by growing AlxGa1-xAs layer on top of GaAs active layer acting as a window material. The AlxGa1-xAs layer is suitably coated with an antireflection material like silicon nitride to quarter-wavelength thickness to minimise reflection losses. Both GaAs and AlxGa1-xAs layers are grown epitaxially. The single layer of GaAs (Zn-doped) has usually a thickness from one to approximately two micrometers. Around 1.2 m thickness, almost 90 per cent of the incident light gets absorbed. Figure 6.4 shows NORMAL PHOTOCATHODE EA

C F

NEA PHOTOCATHODE Vac C

EG

V V = Top of valence bond F = Fermi level C = Bottom of conduction band V ac = Vacuum level

EG

E Ae f f

F V E G = Band gap energy E A = Electron affinity E Ae ff = Effective electron affinity

Figure 6.3. Band model of NEA and normal semiconductor photocathodes.

EA

Vac

Photoemission LIGHT IN

ANTIREFLECTION COATED GLASS FACEPLATE QUARTER-WAVELENGTH ANTI REFLECTION COATING AlxGa1-XAs WINDOW LAYER

NiCr METALLIZING

GaAs ACTIVE LAYER

ELECTRONS OUT

SURFACE TREATED WITH CAESIUM AND OXYGEN

Figure 6.4. GaAs photocathode in transmission mode

the structure of such an assembled photocathode[3]. Activation of the GaAs layer with caesium and oxygen is carried out in ultra high vacuum systems after heat cleaning of its surface. 6.3.5

Transferred Electron (field-assisted) Photocathodes The NEA photocathodes are limited to a band gap corresponding to a cutoff of around 1.1 m. Efforts to push the cutoff wavelength led to the evolution of transferred electron photo-emission. The transferred electron type of photoemitter depends on transferring the photogenerated electrons from a lower to upper valley by means of an electric field (as in Gunn effect) from where they can be emitted. A number of such external fieldassisted photoemission geometries were proposed and studied. One such structure was made of p-type InP on which a thin layer of Ag formed a Schottky barrier. If a reverse bias is applied to the Schottky barrier then the photogenerated electrons are accelerated towards the surface. The accelerated electrons may be placed in L to X valley from which it can penetrate through the metal (Ag) which is activated with Cs/O to lower its electron affinity to 1 eV. InGaAsP materials with a band gap 0.85 eV matched to InP can give lower threshold energy with a Ag layer activated with Cs2O. Since there is no band gap limitations in NEA, the cutoff wavelength was increased with the reduction of band gap and a threshold near 1.5 m was obtained. Later a transmission mode structure giving photoemission up to 2.1 m was achieved by using a In0.77Ga0.23 As layer of band gap of 0.52 eV as the emitter layer. Since

85

An Introduction to Night Vision Technology

this is not lattice-matched to InP, a layer of InAsxP1-x was used in-between. In this configuration, the photons are incident on the back surface (transmission mode) but only those photons for which 0.83>h>0.52 are absorbed in the InGaAs layer. Photons for which 1.35>h>0.83 are absorbed in InAsP layer which has a graded composition meant to give an accelerating field to the photoelectron towards the surface. Photons with energy more than 1.35 eV would be absorbed in InP. The silver film thickness was about 50 A° which was activated to Cs and oxygen. The ultimate limit of band gap for this material system should give the largest threshold up to 3.54 m for InAs. But the main problem of operation of these cathodes is the dark current. Even with cooling to –100°C, the dark current is very high (10 –8A/cm2) and rises sharply with increase of bias voltage. The rapid rise is due to impact ionisation. These cathodes should become quite useful for certain types of applications if the dark current could be reduced. Figure 6.5 graphically illustrates the comparative yield for some of the common photocathodes used[4]. 10 0 5 X 10–1

1. Ag-O-Cs

(S-1)

2. Cs3Sb

(S-11)

3. Bi-Ag-O-Cs

(S-10)

4. Na2 KSb (Cs) (S-20)

2 X 10–1 QUANTUM YIELD (ELECTRON/PHOTON)

86

5. GaAs (Cs,O)

(NEA)

1 X 10–1 5 X 10–2

2 X 10–2 1 X 10–2

5

5 X 10–3

2 X 10–3

400

2 500

600

3

4

900 800 700 WAVELENGTH (nm)

1 1000

1100

Figure 6.5. Comparative quantum yield for some of the common photocathodes used.

Photoemission

6.4

PHOTOCATHODE RESPONSE TIME Response time depends on the materials used. Metal photocathodes have the fastest response time of the order of 10–15 to 10–14 seconds though their quantum yield is the poorest and of an order of 10–4 electrons per photon. It is well understood that as in metals, powerful electron-electron mechanisms dominate leading to relatively restricted release of photoelectrons from near the surface explaining their fast response. Further, electrons from the depth of the material have hardly any chance of escape. Semiconductors with small but positive electron affinities have response times of the order of 10–13 to 10–12 seconds, with yields of 0.05 to 0.25 electrons per photon. Photo excited electrons in these materials do not undergo electron-electron scattering near threshold since a finite band gap separates the highest states filled with a large number of electrons and the lowest conduction band states, enabling longer escape depth. Photoelectrons will still undergo electron-photon (lattice) scattering changing their direction in the bulk, somewhat reducing their energy and increasing their path length to the surface and thereby the increased response time. The highest yields 0.1 to 0.6 electrons/ photons and longest response of the order of 10–10 to 10–8 seconds are obtained in negative electron affinity photocathodes. In these materials even for photons near the band gap, the diffusion length of the thermalized electron is greater than the optical absorption depth. As a result the yield rises much more rapidly than for the earlier type of photocathodes near threshold. The transferred electron (field-assisted) cathodes give faster responses of the order of ~10–11 seconds in comparison to other categories, except metals, and have a quantum yield better than 10–2 electrons per photon. The future interest is in their longer cut-off wavelength values, with a possibility of better quantum yield and still faster response times (Fig. 6.2). The cathode represented by the dot in the figure has a cut-off wavelength at 1400 nm. The two arrows show the development – directions that may be possible. 6.5 PHOTOCATHODE SENSITIVITY In practice, photocathodes have to be used in semitransparent mode so that they can be used essentially in all types of image intensifier tubes. The overall characteristics thus presented are those of the photocathodes in combination with the supporting material. At low wavelength end, one may use a lithium fluoride window with a cutoff at 104 nm though for the visible and near infrared the window material may be lime or boro-silicate crown glass. Fused silica has also been used. Input

87

88

An Introduction to Night Vision Technology

window could also be a fibre-optics bundle supporting the photocathode material. Evaluation is preferred in terms of ‘sensitivity’. Sensitivity is expressed in terms of microamperes per lumen (A/lm) which relates to the luminous sensitivity in white light or in milliamperes per watt (mA/W) as the radiant sensitivity at a given wavelength. Both values are measured using a tungsten lamp with a colour temperature of 2856° ± 50°K as the raw light source. Filters are used to determine the radiant sensitivity at specified wavelengths. Commercial specifications usually give the luminous sensitivity values, as also radiant sensitivity values at around 800 and 850 m for all modern photocathodes. Earlier photocathodes barely came up to a luminous sensitivity of 50 A/lm or to a peak radiant sensitivity of the order of 10 mA/W. Significant increase in these values took place with the introduction of S-10 (Bi-Ag-O-Cs), S-11 (Cs3Sb) and ultimately the photocathodes of choice for passive night vision, i.e., S-20 (Na2KSb(Cs)) and S-25/ERMA. These multialkali antimonides (cessiated) photocathodes now offer luminous photocathode sensitivities of the order of 400 A/lm and radiant sensitivities of the order of 40-45 A/W in between 800-850 nm. Some suppliers of image intensifier tubes claim even higher values of luminous and radiant sensitivities for this family of photocathodes. Subsequent introduction of NEA photocathodes has resulted in still higher values for the sensitivities. Thus the most used NEA photocathode GaAs (Cs,O) has a typical luminous sensitivity of 1300 A/lm and a radiant sensitivity exceeding 50 mA/W. Absolute sensitivity values in mA/Watt against wavelength are plotted in Fig. 6.6 for a number of photocathodes of interest[5]. Spectral response curves in the figure are for combination of photocathodes and windows. Thus, lime or borosilicate glass windows are used in respect of S-10, S-11 and ERMA (Extended red multialkali), or S-25. It will also be appreciated that both luminous and radiant sensitivity values can vary in the same type of photocathode material depending on the processing techniques which may be resorted to by different manufacturers to satisfy their requirements for an end product. The material composition may also be somewhat altered to permit a higher or lower wavelength cut-off and to improve sensitivity for a specific region. The exact values will hence have to be known or determined in each type of photocathode that may be used in an image intensifier tube. Manufacturers of these tubes usually mention in their specifications the photocathode sensitivity in the white light (tungsten source) in A/lm and

Photoemission 1. Ag-O-Cs 2. Cs3Sb 3. Bi-Ag-O-Cs 4. Na2 KSb (Cs) 5. ERMA/S-25 6. GaAs

80 60

ABSOLUTE SENSITIVITY (mA/W)

40

(S-1) (S-11) (S-10) (S-20)

6

20

10 8 6 4 5

4

2 2

3

0.8 0.6

1

0.4 100

200

300

400

500 600 700 800 WAVELENGTH (nm)

900

1000

1100

Figure 6.6. Absolute sensitivity in mA/W vs wavelength

radiant sensitivities in mA/W at specified wavelengths of 800 and 850 nm or at any intermediate wavelength[6]. The higher the radiant sensitivity in the near infrared and better the luminous sensitivity, the more is the response to the night sky subject only to the overall noise limitations in the intensifier system. Analytically, luminous sensitivity is given by the expression 2

 S (  ) E (  ) d

1

0.76

680  y (  ) E (  ) d 0.40

A / lm

(6.7)

where S () is the spectral responsivity of the sensor within its spectral limits 1 and2 in amperes/watt, and E() is the spectral radiance due to the source in watts per square meter. y( ) is the

89

90

An Introduction to Night Vision Technology

relative photospectral response of the human eye, and wavelength is in m (=1000 nm). The quantum yield Q () can also be correlated by the equation: 1.24 S (  ) Q(  )  electrons / photons (6.8)



6.6

DARK CURRENT IN PHOTOCATHODES The ultimate limit to any imaging system’s ability is the photon-to-electron conversion noise, resulting in dark current. This noise for a uniformly lighted area of a photoemitter results from a random release of charge carriers and is hence analytically measurable in terms of a root mean square (rms) photocurrent irms given by

irms  (2ei f )1/2 A

(6.9)

where e is the charge of an electron and f is the measuring bandwidth in Hertz. This dark current in a photocathode arises out of thermionic emission and has a characteristic peculiar to each photo surface. The value is usually greater for red sensitive tubes. The value does increase with increase in sensitivity of the photocathode as also with increase in temperatures under which it is operated. Cooling does help particularly in red-sensitive photocathodes like S-1. Its value is proportional to the surface area of a photocathode also and hence the unit for practical comparison is in amperes/cm2. Table 6.2 compares the values of dark current for different materials with different photocathode sensitivities and different values of long wavelength threshold. Table 6.2 Dark current values for different photocathode materials Material

Type

Max quantum yield (peak wavelength in nm)

Ag-O-Cs Cs3Sb

S1 S 11

0.5(800) 20 (400)

1100 650

Bi-Ag-O.Cs

S 10

10 (450)

Na2K Sb (Cs) ~100 m thick

S 20

25 (400)

Na2K Sb (Cs) ~1000 m thick) Ga As(Cs, O)

S 25 NEA

Other suitable

Long wavelength threshold (nm)

Luminous sensitivity (A/lm)

Dark current (A/cm2)

60 80

10–11 10–15

750

80

10–14

850

~300

10–16

30 (400)

900

~400

10–16

40 (over a wide range) ~3540

950

~1300

10–14 10–8 (for InAs

Photoemission semiconductors

at –100 °C)

In practice for image intensifier tubes, this measurement is incorporated in the definition of equivalent background input. This implies a measurement procedure for screen brightness when the operating potential has been applied to the assembly and no radiation is incident on the photocathode. This value in the device as a whole may be due to field emission, gas ionization, inter-electrode leakage, residual radioactivity and a host of other causes beside the photocathode thermionic emission. However, in all well designed and appropriately assembled image intensifier tubes, these sources of dark current are virtually eliminated except that due to photocathode thermionic emission. This measurement therefore indicates the level of the photocathode dark current. 6.7

SUMMARY The physics of the semiconductors has proved very effective in the understanding of existing photoemissive materials and is helping in the search for better and better materials. The exact processing details including the methodology of depositing material layers are quite important for repetitive production to meet exacting standards for image intensifier tubes. REFERENCES 1. Spicer,W.E. & Gomez, A H., "Modern Theory and Applications of Photocathodes". Photodetectors and Power Meters. SPIE. vol. 2022, (1993), pp.18-33. 2. Sommer, A.H., "Brief History of Photoemissive Materials". Photodetectors and Power Meters. SPIE. vol. 2022, (1993), pp. 2-17. 3. Csorba, I.P., Current Status and Performance Characteristics of Night Vision Aids. Opto-electronic Imaging, (Tata-McGraw Hill Publishing Co. Ltd. 1985). 4. Sommer, A.H. Photoemissive Materials. (New York, London, Sydney, Toronto: John Wiley & Sons Inc., 1968). 5. Walter, G.D.(Ed). Handbook of Optics. (McGraw Hill Company). pp. 4-23. 6. Biberman, L.M. & Nudelman, S. Photoelectric Imaging Devices Vol. 1 & 2. (New York, London: Plenum Press, 1971).

91

CHAPTER 7 PHOSPHORS 7.1

INTRODUCTION Luminescence refers to the emission of light by a material induced by an external source of energy. It may be induced by light which after absorption is reradiated in a different waveband, termed as photoluminescence or by the kinetic energy of electrons termed as cathodo-luminescence. It could also be triggered by the incidence of high energy particles, applied electric fields or currents, or chemical reactions. Luminescent technologies by now embrace liquid crystal devices, gas panels and electroluminescent panels besides the well known cathode ray tubes. The success of these tubes is mainly due to high performance level of modern day phosphor materials. The word phosphor literally meaning light bearer refers to luminescent solids, mainly inorganic compounds processed to a microcrystalline form for practical use of their luminescent property. The earliest phosphors used the naturally occuring Zn2SiO4 and CaWO4 as a thin powder on a mica substrate to act as viewing screens. Usually phosphors are in the powder form but they could also be used as thin films. The image intensifier tube screens have borrowed from the phosphor developments for use in cathode ray tubes. The luminescence we are concerned with is the cathodo-luminescence. 7.2

PHOSPHORS Most phosphors are activated by the introduction of an impurity of the order of a few parts per billion. This impurity which activates the phosphor is known as activator, while phosphor crystal itself is known as the host or matrix. The chemical formulae indicate the presence of an activator in the host crystal. Thus one such formula can be ZnS:Cu indicating ZnS as the host and Cu as the activator. In a sulphide phosphor the dopant of a VII-b group element, i.e., halogens (chlorine, bromine, iodine) or a III-b group element (gallium, aluminium) in addition to the activator is referred

94

An Introduction to Night Vision Technology

to as the co-activator. Thus, the complete chemical formula is of the type ZnS:Cu, Al and ZnS:Ag, Cl. The role of the co-activators in ZnS phosphors is to compensate for the excess minus charge caused by an activator. 7.3

LUMINOUS TRANSITIONS IN A PHOSPHOR When an accelerated electron of high energy say 6 KeV or more penetrates an inorganic crystal, a large number of electron and free holes are produced along its path leading to many possibilities for optical transitions. If the crystal is free from impurities, doping, and lattice defects, the free electrons and holes that have been created in the conduction and valence bands may recombine emitting photons whose energy is equivalent to the band gap [Fig. 7.1 (a)]. These emissions have been rarely observed, except in the case of ZnO where the phosphor has been used in flying spot tubes. In actual practice, the phosphor crystals do have lattice defects, incidental impurities and also deliberately introduced activators and co-activators which create a number of energy levels providing a number of recombination paths for the excited electrons and holes at much less band-gap values resulting in emissions within the visible part of the spectrum. Activators produce deep acceptor levels with different depths. Donor levels on the other hand may be introduced by lattice irregularities, incidential impurities and co-activators [Fig. 7.1(b)]. The above explanation is particular to the two well known CRT phosphors ZnS:Cu, Al and ZnS:Ag, Cl. Differences of colour that is green and blue is attributed to deep acceptor levels respectively created by Cu and Ag at 1.25 eV and 0.72 eV. Time and excitation dependent spectra have been observed for these CONDUCTION BAND

DIRECT TRANSITION

a

b

c

VALENCE BAND

d

DONOR LEVELS

e

ACCEPTOR LEVELS

Figure 7.1. Luminescent transition models in phosphors[1]

Phosphors

phosphors. For phosphors, in general, a number of other transition models are also possible, apart from the direct transition corresponding to the band gap of the host material (Fig. 7.1). The direct recombination transition has been marked as a, while b is the recombination transition between a donor and an acceptor; transition c is between the conduction band and a deep acceptor level, while transition d is between a deep donor and the valence band. The transitions occuring in a well localized luminescent centre or a molecular complex of atoms is represented by e wherein the electrons are confined to the same centre before and after the transition. Such centres are in rare-earth or transition metal ions of the type Eu+3, Ce+3, Mn+2. Rare-earth activators give better results with Y or La as hosts. Configuration coordinate models have been proposed for luminescent centres in the generalized configuration diagram shown in Fig. 7.2. The two curves G and E represent the energies of a luminescent centre in the ground (G) and in the excited (E) states against the configurational coordinate. When the centre is in its state of lowest energy, the configuration coordinates assume the value for which energy is a minimum, i.e., point A on the curve G. Since the equilibrium configuration of the interacting atoms is different for the ground and excited states, the two do not correspond on the CURVE E (EXCITED STATE) CURVE G (GROUND STATE)

O C

ENERGY

CO-ORDINATE

B

D

A

CONFIGURATIONAL CO-ORDINATE

Figure 7.2. Luminescent centre – a configurational model

95

96

An Introduction to Night Vision Technology

configurational coordinate axis. The minimum on the curve E is shown by point B. As the absorption of external energy will occur before the ions have time to readjust themselves to the equilibrium of the excited state, the absorption corresponds to the transition AC. The system readjusts itself and dissipates a little of the energy gained by way of heat to reach the minimal equilibrium point B enabling a radiative transfer to the point D, followed by heat dissipation to reach the point A again. If the thermal temperature of the system is high enough, the centre may be stimulated to the position O and relax its energy to the host crystal, transiting to A along the curve OA without radiative transfer. While the energy difference between A and C corresponds to the peak of the absorption spectrum, that between B and D corresponds to the peak of the emission spectrum. This theoretical and analytic approach has led to a better understanding of luminous centres. 7.4

PHOSPHOR MECHANISMS High energy primary electrons incident on a phosphor may suffer elastic and inelastic collisions or penetrate producing a cascade of photons and internal secondary electrons. Those secondary electrons that overcome the work function escape into the vacuum. High energy electrons which undergo elastic scattering are the reflected and back scattered electrons from the surface, while those undergoing inelastic collisions are the re-diffused electrons with some energy loss. The reflected and back scattered electrons cause contrast degradation of a picture if these are absorbed by the neighbouring phosphor elements. The emission of secondary electrons also has a deleterious effect. In case their rate of emission is significant or more than the rate of arrival of primary high energy electrons, it would shift the potential of the insulated phosphor. The negative charging of the phosphor screen by reducing its potential may seriously reduce the light output. This charging is prevented by depositing a thin film of aluminium, which is penetrated by high energy electrons on the surface of the phosphor. All these three factors, i.e., elastic and inelastic collisions, and the escape of secondary electrons lead to the attenuation of the absorbed energy and thereby the efficiency of cathodo-luminescence. Incident higher energy electrons result in a reduced spread of luminescence along their path within the phosphor, as against lower energy primary electrons. Figure 7.3 is illustrative of the observations of the cathodo-luminescence of a crystal excited by a fine electron beam. It shows that with increasing energy of the incident electron beam, the spread in the phosphor is lesser and lesser as the depth of penetration is progressively increased, i.e., an electron with a

Phosphors

lower energy has a larger probability of energy dissipation. At relatively lower incident electron energies the luminescent volume has a hemispherical shape, while at higher energies, the penetration volume is along a narrow channel ultimately terminating into a large spherical volume. The relationship between the energy of the primary electrons and the depth to which it penetrates the phosphor are related by an empirical formula of the type

E  E 0 {1  ( x / R )}1/ 2

(7.1)

where E0 is the energy of the primary electron, E the energy at a depth x and R a characteristic of the material. The energy reduces to zero when x = R. R in turn has been empirically defined in terms of material parameters, such as its bulk density, the molecular weight and the atomic number. Another empirical formula specially for the ZnS phosphors which can be applied up to 20kV is given by

X   116 . E10.65 (in nm)

(7.2)

where X' is the depth where the energy of the primary electrons (E0) drops to e-2 of its original value. One more aspect is that when E0 is decreased, while the beam current is maintained, the cathodo-luminescence drops to LOW ENERGY

MEDIUM ENERGY

FINE INCIDENT ELECTRON BEAM

NARROW CHANNEL

HIGH ENERGY

PHOSPHOR MATERIAL

Figure 7.3. Luminescence spread as a function of incident beam energy (schematic).

97

An Introduction to Night Vision Technology

zero at around 100 V to a few kV depending on the material in the powder form and its method of preparation. Cathodo-luminescence increases linearly for energies above the threshold though at higher values the increase is slower. Figure 7.4 shows the luminescence intensity of ZnS:Ag, Cl against the incident electron energies. The energy efficiency of cathodo-luminescence is obviously a product of the mechanisms of energy transfer that take

2

LUMINESCENCE INTENSITY (mW/cm2)

98

1

0 0

2

4

6 8 VOLTAGE (kV)

10

12

Figure 7.4. Incident electron energy vs luminescence of ZnS: Ag,Cl, phosphor.

place inside a phosphor when a high energy electron is incident on it. These mechanisms relate to surface reflection and scattering and to the division of the incident electron energy that enters the phosphor into pair production (electrons and holes) and loss by photon emission. Effective models have been developed which give results very close to the practical values. Thus for ZnS:Ag, Cl the observed value of efficiency 0.21 compares favourably with the maximum possible theoretical efficiency of 0.28. The efficiency also depends on the current density of the incident electron beam and the

Phosphors

temperature. It is reduced with increase in current density as also with increase in temperature. The former results in brightness saturation while the latter leads to thermal quenching. 7.5

REDUCTION OF LUMINESCENCE EFFICIENCY Luminescence efficiency is quenched by an increase in temperature, level of excitation, presence of undesirable impurities and high activator concentration. Increase in temperature generally leads to an increase in non-radiative energy transfer thus reducing the luminescence efficiency. See the path OA in Fig. 7.2 for a luminescent centre. Brightness saturation is possible as a result of a high level of excitation. Mechanisms of brightness saturation are not fully understood though a number of explanations seem to explain this behaviour partially. Undesirable impurities act as killers of the luminescence. Thus the presence of a few parts of Fe per billion in a ZnS phosphor may kill its luminescence totally. One of the reasons is that the impurity centres capture the free carriers in competition to the luminescent centres and enable a non-radiative transfer of energy. Resonance energy transfer is also possible from a nearby luminescent centre. When the concentration of an activator is too high, a fraction of the activators behave as killers and induce quenching. 7.6

LUMINESCENCE DECAY It is observed that after an electron beam ceases to fall on a phosphor causing luminescence, an afterglow persists for some time. This time is known to vary from 10–8 seconds, i.e., of the order of a spontaneous emission to a few tenths of a second or longer. As the response time of the human eye is around 0.1 seconds, it is obvious that the decay time more than 0.1 seconds would be registered by the human brain. This delayed luminescence is called phosphorescence while the one not registered, i.e., below 0.1 seconds decay time is referred to as fluorescence. As the decay time of the luminance in the case of sulphide phosphors is mainly dependent on the time spent by carriers in the luminescent centres which does not exceed 10–1 seconds, these phosphors are usually not phosphorescent. One observes that the decay in these phosphors follows a time power law of the form

I t  I o (1  At ) n  I o At  n

(7.3)

Where It is intensity after a time t after termination of excitement, Io the intensity under excitement, and A a constant. The exponent n value could be 1.1 to 1.3 according to a number of workers in the field. A defect or an impurity which allows a charge carrier to remain

99

100

An Introduction to Night Vision Technology

for a while before these reach the luminescent centres give rise to trapping levels and lead to phosphorescence, i.e., an afterglow lasting for more than 0.1 seconds (Fig. 7.5). The decay is prolonged by the time the charged carrier spends in the traps. This time would be dependent on the depth of the trapping centre in relation to the conduction band and temperature, and would be inversely proportional to the probability of non-radiative transfer between these levels. It has also been reported that in some phosphors the decay is strongly dependent on the duration of excitation, for example, ranging from microseconds for short excitation to milliseconds for longer excitation. Steady-state values are reached for longer exposures. In practice, the specified decay value in a phosphor or a mix of phosphors has to be such that it does not cause scintillations due to fast decay and at the same time it does not cause multiple images of a moving object resulting from a slow decay. 7.7

PHOSPHOR APPLICATIONS Phosphors have found a large number of commercial applications ranging from television screens to vacuum fluorescent

CONDUCTION BAND (EMPTY) TRAPPING LEVELS

RADIATIVE TRANSFER

ACTIVATOR LEVELS VALENCY BAND (FILLED)

Figure 7.5. Luminescent process with trapping levels

displays. Additive mixing of blue, green and red phosphors emissions allows realization of colours within the chromaticity diagram, and thus appropriate phosphor screens for the colour TV displays. Cathode ray tubes (CRT’s) form the basic unit for a large number of applications including scientific and technological, such as terminal displays, projection displays, beam index tubes,

Phosphors

flying spot scanners, radar tubes, storage tubes and the like wherein the selection of a phosphor or phosphors would depend on the requirement and be decided in terms of phosphor grain size, phosphor thickness, nature of emission and the decay time or persistence of vision. Phosphors for applications like image intensifier tubes, or electron microscopes call for high resolution phosphors. The grain size has to be small to reproduce images with high resolution. The size cannot be reduced much as it results in the decrease of the luminous efficiency. The minimum size is restricted practically to 2 m. Green emitting phosphors are generally preferred for direct visual observation because of their spectral match to the photopic human eye. Blue-emitting phosphors are in use for photographic recording because of their good spectral match to the silver-halide photographic films. 7.8

PHOSPHOR SCREENS CRT screens usually have a phosphor weight of about 3–7 mg/cm2 on its glass surface. The phosphor particles may be of size 3–12 m and 2–4 particle-layer thick. The aim is to maximize the emission intensity vis-a-vis its optical screen weight. Image intensifier screens are usually built up on the fibre-optics windows of the tube systems. Both in the case of CRT’s and the fibre-optics windows for the I.I. tubes, the side on which the electron beam impinges is coated with a thin aluminium film. The film works as an electrode which prevents the screen from negative charging during excitation and thus increases the output. Further, it also prevents the light generated in the screen to feedback to the cathode and reflect the light to increase its effective output. Applied voltages have to be relatively higher to penetrate the thickness of this aluminium film. Thus, around 3 kV is the minimum estimated value for penetration through an aluminium film of around 300 nm thickness. About 30 kV applied voltage is applied for X-ray image intensifiers and somewhat lower, i.e., of an order from 9–16 kV for image intensifiers in the optical region. The screen thickness in the case of phosphors for I.I. tubes may be of the order of 100 nm. Usually the green emitting phosphor ZnS:Cu, Al phosphor may be preferred with a particle size of around 2–3 m for image intensifiers though blue emitting phosphor ZnS:Ag has also been referred to. Emission peak of the green phosphor at 530 nm can be shifted to longer wavelengths either by employing a solid solution ZnI-x Cdx S, or by introducing a deeper acceptor level due to gold. Usually the exact parameters of a phosphor or for

101

102

An Introduction to Night Vision Technology

that matter even for a photocathode, for I.I. tubes, like the chemical formulation, layer thickness and its method of application or deposition is an information that each manufacturer keeps to himself. The broad specifications of an user for all these products is satisfied for the whole unit, i.e., the image intensifier tube (I.I. tube). Thus for the phosphor the relative spectral response may suggest a peak value at 510-560 nm with a bandwidth of about 200 nm and a response not exceeding 10 per cent of the peak value at 650 nm. The user may also specify that the afterglow should not exceed 1 per cent of intensity after termination of exciting energy corresponding to a low value of input illumination at the photocathode end within one-tenth of a second. Specifications may also be laid down for field emission or scintillations. Such user requirements generally interrelate phosphors, the electron lens, power supply and the photocathode – the main constituents of an I.I. tube, leaving the design and material choice to a prospective manufacturer. Figure 7.6 shows a section through an image intensifier phosphor screen. The fibre-optics faceplate is the substrate for the phosphor. After coating with the phosphor, it is coated with an aluminium layer facing the incident electron beam. This face may be plane or curved depending on the aberration characteristics of the electron beam. In an alternative process the core of the fibre may be selectively etched away to a depth of a few microns before ELECTRON BEAM CURVATURE TO SUIT ELECTRON OPTICS

ALUMINIUM LAYER PHOSPHOR

FIBRE CLADDING FIBRE CORE

LIGHT OUTPUT

Figure 7.6. A section through a phosphor screen for I.I. tubes

Phosphors

depositing the phosphor. This confines the light on excitement within the channel and helps to reduce the cross-talk. For use in I.I. tubes, phosphors require to have a high luminous efficiency (in terms of lumens per watt) as also an optimum rendering of contrast and resolution in relation to the other components of the system, i.e., photocathode, electron lens, micro-channel plates, etc., that may have been used in the system. While luminous efficiency is a property of the phosphor material, its optimum thickness, larger grain size and the method of its deposition; the imaging properties would be more related to its smaller grain size and optimization with the fibre-optical components, i.e., micro-channel plate and faceplate. The latter properties can be evaluated in terms of modulation-transfer function (MTF). 7.9

SCREEN FABRICATION

After purification of the raw material and removal of the killer materials, the constituent phosphor compound is synthesized. After synthesis, crystal growth is brought about by firing. Next, the coagulated phosphor grains are suitably milled and dispersed uniformly to form a liquid slurry. Fine particles for high resolution of the order of 2–5 m are separated from the larger grains by the sedimentation process. The dispersal and adhesion of the phosphor particles should be suitable for the technique that may be adopted for screen fabrication. Various processes for screen fabrication include settling under gravity, brushing and dusting techniques. Electrophoretic method has been suggested to obtain dense monochromatic phosphor screens with fine particles that are required for high resolution applications as in I.I. tubes. This method is preferred, as the migration of well dispersed fine powders is more affected by the applied electric field than by settling due to gravity. Further, the deposition due to created electric field is such that the pinholes if formed attract phosphor particles preferentially. The result is a dense uniform screen free of pinholes with a smooth surface. Brushing technique has also been preferred to the settling technique as it results in significantly better MTF value for the I.I. tubes phosphor screens though with a slight decrease in luminous efficiency. Though as referred to elsewhere, the exact type of phosphor, its thickness and the nature of screen fabrication is a company’s guarded information, literature does refer to sulphide P.20 and RCA 1052 phosphors with a peak frequency of 560 nm for use in wafer I.I. tubes. Likewise a mixture of silicates, P1 and P39 is also known to have been used in second generation 25 mm

103

104

An Introduction to Night Vision Technology

electrostatically focussed invertor tubes. The efficiencies are stated to be around 15 lm/W for 6 KeV electrons. The techniques of making screens with appropriate phosphors or phosphor mixtures continue to be refined for better and better contrast and resolution in imagery and for more accurate colour rendition. Thin film technique has also been introduced. Efforts continue to develop phosphor materials with reduced brightness saturation, special colour characteristics, better ageing properties and control over persistence. 7.10

PHOSPHOR AGEING Phosphors are known to age under use for sometime, and some portions in the screen may thus turn brown or black. In case it happens immediately or in a short duration of exposure to the electron beam, it is referred to as burning. The same effect can take place over a reasonably long time when the term ageing is more relevant. The substrate glass may also get affected turning the effected portions to brown colour. This is known as browning. Ageing is generally dependent on the charge per unit area falling on the phosphor. Sometimes the phosphor darkening can be overcome by thermal bleaching, that is the phosphor may be annealed at a few hundred degrees Celsius. According to Leverenz, the harder, high melting, water insoluble materials are most resistant to loss of luminescence during operation. Browning of glass would certainly be enhanced by a relatively poor packing of grains leaving a large number of pinholes. An appropriate technique for phosphor deposition is therefore an important parameter from this point of view also. Burn-in profile tests are laid down by the users for the I.I. tube as a whole. These are in the nature of a large number of cyclic operations for minimal and maximal values of the luminous gain against time. Screens can be examined with high power magnifiers at both high and low levels of illumination. REFERENCES 1.

2.

Flynt, W.E. "Characterization of some CRT Phosphors", Ultrahigh Speed and High Speed Photography, Photonics and Videography. Proc. SPIE. 1989, vol. 1155, pp. 123-30. Hase, T., et al. Phosphor Materials for Cathode Ray Tubes", Advances in Electronics and Electron Physics. (Academic Press. Inc., 1990) p. 79.

CHAPTER 8 IMAGE INTENSIFIER TUBES 8.1

INTRODUCTION An image intensifier tube essentially accepts a photon spread from a quantum starved scene below the visibility level through an optical system on its photocathode. Such photons release weak electrons which in turn are accelerated through an electronlens system and made to impinge on a phosphor maintaining correspondence between the optical photon-spread on the photocathode and the amplified optical output from the phosphor. This amplified output from the phosphor can be coupled to an eyepiece system for direct vision or to a video system for vision on a monitor. Thus if h1 is the energy of the incident photon on a photocathode and h2 is the energy of the output photon corresponding to the electrons impinging on the phosphor, one could indicate this double conversion as h1 (on photocathode) ———————>e– e– (accelerated) ——————————>h2 (from phosphor)

The range of 1 which release electrons from the photocathode depends on its spectral sensitivity. Likewise, the limits of 2 are defined by the spectral sensitivity of the phosphor. These aspects have been well discussed in Chapters 6 and 7. The original photon-spread focused on the photocathode is formed by suitable optical systems as discussed in Chapter 5. Further, in modern image intensifiers, the accelerated electrons are significantly multiplied to increase the number of impinging electrons on a corresponding area of the phosphor through the use of micro-channel plates. Historically, image intensifier tubes (I.I. tubes) have now been classified in terms of generations based on the type of photocathode that has been utilized. Thus, in the 1940s, the zero generation made its first appearance using the S-1 photocathode, wherein artificial illumination in the near infrared beyond the visible

106

An Introduction to Night Vision Technology

range was a definite requirement for its proper functioning. The systems matured after the next two decades or so and were reasonably operative in the night environment till research led to better photocathodes and to sensor development for detection of light beyond the visible range. Interest to these systems was particularly drawn when it came to be known that Russian tanks could move about freely during the nights without any lights in the then East Germany. The World over, armies built these systems which were soon to become obsolete on the advent of better photocathodes. These systems however dominated the early sixties and were built also in India virtually parallelly to those in the more advanced countries of the West. Generation-1 tubes also started making their appearance in sixties based on alkali photocathodes at sensitivities around 200 A/lumen which could be later cascaded with each other through fibre-optic input and output windows to enable reasonably higher gains. The systems built around these tubes had no need for any supporting artificial illumination as in the case of Generation-0. These tubes are known as Generation-1 I.I. tubes. This approach proved to be quite spectacular at the time of its introduction and research activities were thus directed to the development of better and better photocathodes and smarter techniques for amplification. It was soon realised that the photon rate from a night sky incident on a photocathode through a suitable optical system is greater by 5–7 times in the 800–900 nm region as compared to that in the neighbourhood of 500 nm. The output signal could thus be significantly improved, if the photocathode is also red-sensitive. This brought in more sensitive S-25 or ERMA (extended red multi alkali photocathodes) for use in I.I. tubes in preference to the standard S-20. This development coupled with the technological development of micro-channel plates (MCPs) to increase the number and energy of impinging electrons on the phosphor, brought in Generation-2. The military significance was all the more as it not only increased the sensitivity and hence the night vision range of the systems designed around it, but it also drastically reduced the weight as one could now substitute a single diode Generation-2 tube for a threestage Generation-1 with better results. Proximity tubes without an electron-lens but with a MCP compacting it further and with a further reduction of weight could also be produced for a number of applications. Systems based on Generation-2, I.I. tube have been produced in large numbers within the country and these could withstand tough competition from contemporary production of the West. Generation-1 systems produced earlier were also upgraded. Meanwhile a good theoretical understanding of the photocathode physics has led to the development of Negative Electron Affinity photocathodes, further shooting up the sensitivity values to an order of 1000 A/lumen or

Image intensifier tubes

better. Though these photocathodes have slightly lower values for signal-to-noise ratio in comparison to Generation-2 and 1 tubes, their excellent sensitivity to the entire spectrum including the red has led to systems with signal detection at much lower levels of ambient light. I.I. tubes incorporating NEA photocathodes are now called Generation-3. Table 8.1 gives a comparison of I.I. tubes belonging to different generations, as these developed from the earlier days. Table 8.1. The family of I.I. tubes Year Gen

1940s onwards Gen-0

Noise factor

—

1960s onwards Gen-1 1

1970s onwards Gen-2 (1.35–1.7)

1980s Gen-2 (wafer)

Gen-3

(1.3–1.6)

(1.75–2.0)

NEA (Ga-As cessiated)

Photocathode type & constituents

S-1 (Ag-O-Cs)

To begin Usually with S-20 S-25 Multialkali Later:S-25 Extended red multialkali

S-25

Tube characteristic & technologies

Single diode low gain high dark current

(a) Single diode (ceramicmetal seals) (b) 3-diodes in cascade (Fibre-optics faceplates)

(a) Electrostatic inverter with micro-channel plate(MCP) b)Brushing techniques for the Phosphor

Wafer tube; Improved double MCP’s proximity focusing

Sensitivity (A/1m)

30

Progressively with more advanced version from 200 > 300 usually 240

300

300

1000

900

350

90

75

High performance (10–3 Lux)

Lighter tube Very light (High tube performance 10–3 Lux) anti blooming

Weight (gm) Remarks

— Active type (requires illumination)

(a) Risk of blooming (b) Image distortion

(a) Visibility down to 10–4 Lux (b) Strong sensitivity (c) Spectral response from 0.6 m to 0.9 m

107

108

An Introduction to Night Vision Technology

Image quality and resolution at low light levels is determined by photocathode sensitivity, spectral response, spectral emission of the phosphor screen, signal-to-noise ratio of the I.I. tube and its integrated radiant power or luminous gain. As several hundred photons per storage time of eye are needed to experience a comfortable visual sensation, I.I. tube amplification should be able to cover up losses caused by low quantum efficiency of the eye and the low transfer efficiency of light transferred from the phosphor. As this percentage may be around 1, it would mean that around 105 photons should be produced per photoelectron to cause a visual sensation. The ultimate resolution is however determined by the statistics of the photoelectrons released from the photocathode and amplified during the storage time of the eye. As the P-20 phosphor matches the spectral distribution of the photopic eye, it is preferred as the phosphor of choice in I.I. tubes. Another phosphor 10-52 which is somewhat closer to mesopic response of the eye is also in use. 8.2

FIBRE OPTICS IN IMAGE INTENSIFIERS Generation-1 single I.I. tubes require to be cascaded for better overall gain, so as to be useful under low light level conditions. Various approaches were therefore tried including insertion of phosphor photocathode dynodes. These dynodes consisted of a thin plate of mica or glass with a phosphor layer on one surface and a photocathode layer on the other mounted in a single glass tube envelope. However, reasonable imagery, i.e., with freedom from curvature of the electron image field and radial distortion was possible only by introducing magnetic focusing between the flat dynodes resulting in a cumbersome and expensive design (Fig. 8.1). A similar approach using secondary emission multiplier dynodes also did not give any better overall results[1]. Simple economical and modular cascading became possible only after the appearance PHOTOCATHODE GLASS ENVELOPE PHOSPHOR PHOTOCATHODE

PHOSPHOR

THIN SHEET OF GLASS OR MICA (DYNODE)

Figure 8.1. Principle of earlier cascaded image-intensifiers

Image intensifier tubes

of the fibre-optics fused faceplates and their use as input and output windows. Later, Generation-2 tubes became a success due to the introduction of micro-channel plates. Introduction of fibre-optic twisters in the proximity I.I. tubes of Generation-2 was a further advancement. The contribution of fibre-optics components has therefore been quite important to the continued use of the I.I. tubes for night vision. All the three components, i.e., fibre-optics faceplate, hollow fibre micro-channel plates and fibre-optics twisters continue to be used for some purpose or other either singly or in combination in modern day I.I. tubes. 8.2.1

Concepts of Fibre-optics Though it is not possible to deal with the subject of FibreOptics in detail within the confines of this volume, a brief introduction to understand some of the concepts relevant to the functioning of fibre-optical components for use in I.I. tubes may be necessary[2]. Conduction of light along cylinders by multiple total internal reflections has been known for quite sometime. However, it was only in early fifties when glass-coated glass fibres made their appearance, that there was a technological quantum jump. Earlier uncoated fibre in air used to get contaminated very easily and did not provide a proper interface for multiple total internal reflections. Techniques of fabrication of multiple-fibres subsequently led to the successful manufacture of the fused fibre-optics faceplates. The term Fibre-Optics was first introduced by Kapany. Figure 8.2 shows the path of an optical ray through a glass-coated glass fibre. Rays after refraction from the entrance face strike at the interface of the core and the cladding. All the rays which strike at the interface at an angle equal to or greater than the critical angle get trapped within the core of the fibre and are thus transmitted to the exit-end. ANGLE OF REFRACTION CORRESPONDING TO MAXIMUM ACCEPTANCE ANGLE  cr na

ncl 

C

nc

MAX. ACCEPTANCE ANGLE

Figure 8.2. Path of a light ray through a optical fibre

109

110

An Introduction to Night Vision Technology

Important aspects of relevance to fibre-optical components, particularly the faceplates are (i) the numerical aperture and (ii) the absorption. As the optical beam through the fibre transits a relatively small length, the absorption is not that significant as in the case of optical fibres for signal communication. Nevertheless, the material of the core and the cladding should be exceptionally uniform and absorption as low as possible keeping in view the fact that the refractive indices of the core and cladding will be decided based on the requirements of the numerical aperture. Assuming good quality fibres, i.e., with the minimal of absorption, and with neat interface between the core and the cladding, the total light transmission is dependent on its numerical aperture. This parameter defines the cone of light that is accepted by an optical fibre or for that matter by a lens system at its entrance aperture. It can easily be shown that the numerical aperture(N.A.) of an optical fibre quantitatively defined by na sin is given by



n a sin  n c 1 n cl /n c 



2 1/2

(8.1)

where na is the refractive index of the medium from which the light is incident on the fibre, i.e., air or vacuum, nc is the refractive index of the core of the fibre and nc1 that of the cladding. Angle  is the angle of incidence (Fig. 8.2). Equation 8.1 shows that the na will tend to be a maximum if the core refractive index is higher and the cladding index is lower. Maximizing this value enables greater acceptance angle of the incident beam. This angle can be maximized to 90° with suitable selection of refractive index values for the core and the cladding. na can be unity. In other words, the optical fibre can transmit all the light that is incident on it, which is not quite true of an optical system. To attain this sort of working from an optical system a lens system will be required with a numerical aperture of F/0.5! It will be observed that the factor ncl /nc is also sine-inverse of the critical angle cr at the interface of the core and the cladding. If the angle of refraction at the entrance face is c then if angle (90-c) is equal to or greater than the critical angle cr the ray will remain trapped within the core and undergo multiple reflections till it reappears at the exit end (Fig. 8.2). The other aspect is that the light incident on an optical fibre received all over its maximum acceptance angle is somewhat averaged out by multiple reflections by the time it reaches the output end. 8.2.2

Fibre-optics Faceplates If a large number of such optical fibres are packed together parallely over a short distance of the order of a few mm,

Image intensifier tubes

the result is an optical fibre faceplate which is in the form of a disc. The disc size is determined by the size of I.I. tubes that it has to fit into. The standard sizes usually are 18, 25 and 40 mm in diameter. Both the end faces of the disc are polished. The packing should be efficient so that maximum light incident is on the core. A hexagonal shape for individual fibres seems to be preferable. Additionally one has to ensure that the light incident on the cladding and which may leak from the core into the cladding because of incidence at angles greater than the acceptance angle is absorbed. For this purpose, strategically placed black glass rods known as extramural absorbers are also introduced in the pack (Fig. 8.3). EXTRAMURAL ABSORBER

Figure 8.3. Likely placement of extramural absorbers

Such a faceplate can be optically characterized in terms of the optical resolution that it may offer and in terms of MTF as is the case with other optical and electro-optical subsystems. Thus, if an image is formed on one of the end faces of a fibre-optics faceplate, the intensity pattern is faithfully carried through the fibre to other end but with a resolution corresponding to that of the corefibre diameter, or rather centre-to-centre distance of the adjacent fibres in the fibre pattern. Thus, to have better resolution the fibre diameters should be lesser and lesser. However, diffraction considerations limit this diameter to an order 5 m or so. Figure 8.4 shows a typical fibre faceplate as designed and fabricated at one of the laboratories in India. Fibre faceplates can also be used as field flatteners in optical systems to correct for the curvature of field where other aberrations are under reasonable control. This also applies to electrostatic lens systems which are in use in image intensifiers particularly for cascading (Fig. 8.5). It will be observed that to

111

An Introduction to Night Vision Technology

Figure 8.4. View through a fibre faceplate

simplify the electron-lens systems, both the input and output fibre-optics faceplates have been suitably curved on the inside surfaces facing the vacuum. This helps the design to be both modular and simpler. In fact, the success of the three tube Generation-1, image intensifiers resulting in a high luminous amplification was entirely due to fibre-optic faceplates. The output from the Generation-1 first diode output faceplate could be coupled to the input window of the second diode, and likewise, the output from the second diode could be coupled to the third one (Fig. 8.5). A large number of optical, thermal, and chemical criteria must be satisfied in the fabrication of a usable fibre-optics fused plate. One of the important methods for fabrication of such fibres is the rod-in-tube process. A high numerical aperture fibre results by drawing high-index glass rods snug-fit in tubes of low refractive index. The glasses have a high degree of homogeneity and are free Fibre input window

Fibre output window

Fibre input window

Fibre output Fibre input window window

Final output window

50 mm

112

180 mm

Figure 8.5. A sectional view through a Gen-1 cascade system

Image intensifier tubes

from bubbles and seeds. The glass types are also so chosen that they are compatible both thermally and chemically. Obviously, both the rods and tubes must have thoroughly clean and smooth surfaces before being snug-fit and placed in a drawing machine. The drawn fibre is dipped through a dark solution of a ceramic material which provides an absorption coating also known as extramural absorption coating or EMA. To ensure precise diameters of the output fibres, the thermal gradient in the furnaces, the rate of sliding the rod-in-tube combination into the furnace, and the rate of drawing the output single fibre are controlled very critically and effectively. A proper calibration of the drawing and furnace equipment is essential before good results can be expected. The nominal diameter of the output single fibre is not allowed to be varied by more than a few per cent of its value to maintain excellent uniformity. The nominal value of single fibres may be from 0.5 mm to around 3 mm. Exact diameter is decided by the nature of the materials and the equipment that has been used, as also the equipment that will be used to produce multiple fibres. The single fibre is usually cut in short lengths say of the order 250 mm or more. These cut single fibres are then grouped and aligned in graphite moulds of usually hexagonal or square crosssection. Alignment is fully assured manually or through utilization of appropriate jigs and fixtures. This mould is next raised to a temperature corresponding to the softening point of the fibre coating material to accomplish tacking between the single fibres. This group of single fibres is then redrawn after appropriate annealing resulting in multiple fibres using the same or similar drawing and furnace equipment. The drawn multifibres are cut to right lengths and aligned in a suitable jig. High quality fusion between the multiples is ensured by controlled heating to the softening temperature of the coating material and by appropriate pressure. This is followed by annealing to eliminate strain or inhomogeneities in the composite. The boule so formed can be sliced in appropriate thickness to form the component fibre-optic plates. Both surfaces of a disc or plate so available need to be polished and surfaced, as per the requirements to form suitable faceplates. Needless to say, control and testing has to be adopted at each stage for optical and mechanical control with precise instrumentation besides ensuring complete vacuum tightness. Degree of cleanliness while drawing, fusing, sawing, surfacing and polishing has also to be of a very high order so as to obtain maximum efficiency from the finished product. It has also to be ensured that the materials used for the

113

114

An Introduction to Night Vision Technology

fibres and coating are not such as to poison photocathode or phosphor materials. The fused plate as a whole should also have a compatible thermal behaviour for correct sealing to the envelopematerial of the photoelectric device in which it is to be lodged, to avoid the possibility of bad sealing or development of cracks. 8.2.3

Micro-channel Plates A micro-channel plate (MCP) is usually a disc-shaped compactly packed assembly of short length micro-channels finished flush with the two plane end faces of the disc. These two faces may be parallel to each other or have a wedge. Each micro-channel is basically a capillary or more correctly a hollow fibre. The material of the hollow fibre has a certain amount of electrical conductivity and hence each micro-channel may be considered as a continuous dynode electron multiplier. Its introduction in a I.I. tube enables high gain with a minimum size and weight with an additional advantage, as the saturation characteristics of each channel limit the blooming effect over itself and restricts its spread to the nearby area. Usually the channel diameter is of an order of 10-15 m with the overall disc thickness of a millimetre or less. The diameter-tolength ratio of the micro-channels is dictated by the electronmultiplication considerations and the material of the capillary with a view to obtain as linear a gain as possible with minimum noise. Amplification fluctuation should be minimum both with time for each micro-channel and of one micro-channel with respect to another. Thus, uniformity of channel material as also the diameter and lengths of each micro-channel have to be critically controlled. The microchannel diameter is also related to the desired spatial optical resolution. As shown in Fig. 8.6, an electron entering any of the individual channel is reflected from the channel walls releasing secondary electrons. Now if a voltage is applied between the input and output faces of MCP providing a potential gradient, then these electrons get further accelerated and result in many more electrons as a field exists accelerating these secondaries. This process of generation of secondary electrons continues till the output end. The gain in each micro-channel of MCP (assuming similarity in all the micro-channels) is dependent on the average number of collisions of the electrons within the channel walls and on the emission coefficients on each collision[3]. The value of the coefficient is always a maximum for the first collision. For subsequent collisions, this value goes on decreasing. If the value of the first collision is denoted by t and the average secondary coefficient by s, it can be shown that the gain (g) is approximately given by

Image intensifier tubes

A MICRO-CHANNEL PLATE

ELECTRON AVALANCHE

BIAS ANGLE

MICRO-CHANNEL VOLTAGE (GAIN CONTROL)

10-15  ION TRAP SECONDARY ELECTRONS

INPUT ELECTRONS

MICRO-CHANNEL ~ 500 

OUTPUT ELECTRONS (G x INPUT ELECTRONS)

Figure 8.6. Electron amplification through a micro-channel

g ~ t .s N

(8.2)

Where N is the total number of collisions that take place. For a given diameter, the number of collisions is dependent on the direction of the incident electron and the length of the micro-channel. Noting that the diameter of the micro-channel has to have an optimum value from the point of view of optical resolution or MTF, the parameters that can be varied to improve on the gain are: (a)

Increasing the potential gradient to further accelerate the electrons in the channel thereby increasing the s value.

(b)

Increasing the value of N, i.e., number of collisions. This suggests an increase in the length-to-diameter (l/d) ratio and accommodating steeper direction for the incident electrons.

It has been stated that for MCPs with 15 -m centre-to-centre hollow fibres, the gain roughly doubles for every increase of 50 V.

115

116

An Introduction to Night Vision Technology

Having optimised the value of this potential, further gain is possible only by increasing the l/d ratio. Here, one observes that for a given constant potential, the secondary emission coefficient goes on decreasing since the impact energy decreases with each further collision. In other words, s gets a lower and lower value with increase in N tending to reduce the overall gain. Hence, the factor l/d cannot be increased beyond a certain value of N for which g becomes less than 1. To appreciate this behaviour better, closer approximations than in Eqn 8.2 have been adopted. One such is the calculation of the decrease value of s at each collision as a function of the potential gradient. Using better computing methods, it has been possible to calculate the value of l/d for a given potential gradient which maximizes the gain. The gains desired are of an order of 102 to 104 at length/d ratios between 30–40 and operating voltage between 600– 900 V. The other important factor apart from gain is the noise. Fixed pattern noise in an MCP is due to variation in gain with time or between adjacent channels. These result in scintillations or speckles. This sort of output may be observed even in the absence of an input current or as a superposition in the output when an input current is also present. Most of these spurious fluctuations arise from fieldemitted electrons emerging from the uneven topography of the input ends of an MCP. Better control in fabrication technology reduces these effects to within tolerance. Control in diameter is also very essential to a great degree as diameter differences may contribute significantly to the fixed pattern noise. Excellent manufacturing technology also minimizes the dark current and gives it a consistent value. The signal induced noise-figure is a function of the open area of an MCP. Electrons hitting the closed area may produce secondaries which may effect the I.I. tube noise rather then the MCP noise. It is observed that improvements in the noise figure may be brought about by increasing the open area ratio, using materials which enable better values for secondary emission coefficient for the first strike, and decreasing the first strike depth. Thus, micro-channels at the input end have been coated with MgO or CsI. As in an I.I. tube, the electrons arrive at the MCP at normal or near normal to the surface, biasing of the surface can be resorted to for decreasing the first strike depth. A bias of around 10° is given to the face to enable steeper angles of strike at this shorter first strike depth. A higher bias is not useful as it results in distortion in focusing of the output electrons. Curved MCPs have also been introduced in which each channel is curved, so that the first strike depth is decreased.

Image intensifier tubes

MCPs are fabricated out of glass containing basically silica, alkali ions for desired softening and annealing and a requisite amount of lead and bismuth oxides to provide conductivity. Rod-intube method is adopted as it is necessary for making fibre faceplates as detailed in section 8.2. After finishing fused fibre plates in all respects, the core is removed by a chemical etching process. Obviously, the core-material should be easy to remove while the cladding which ultimately forms the micro-channel is unaffected. Selective etching at the input end may be done first to improve on the open area ratio. The channel conductivity is adjusted by appropriate reduction of lead or bismuth oxides by controlled firing in an atmosphere of hydrogen. This activates the MCP. A suitable metal is next evaporated on its front and back surfaces. Techniques are also adopted by suitable coatings to improve on the secondary emission coefficient on first strike and to prevent ion feedback. Alternative techniques are also available by drawing hollow fibres from the beginning. Obviously the control and care in the fabrication of MCP has to be much more than what is necessary in the case of fused fibre-optics faceplates. As indicated in Table 8.1, Generation– 2 I.I. tubes have proved to be a success because of the incorporation of MCPs in their design. 8.2.4

Fibre-Optic Image Inverters/Twisters Image inversion is accomplished in some types of proximity I.I. tubes using a fibre-optic component. This internal component within the tube avoids the need for additional optical components and enables a compact and light weight device. Such components are fused optical fibre bundles in which varying degrees of twists have been imparted during the fabrication process. Images are known to be transmitted without distortion through several complete rotations of the fibre bundles. In a bundle around 13 mm in diameter, a 180° rotation has been achieved in 13 mm length itself. As the image inversion takes place as a result of the twisting of the optical fibres, these components are also referred to as optical twisters or inverters (Fig. 8.14). 8.3

ELECTRON OPTICS Any inhomogeneous axis-symmetrical electrical or magnetic field acts upon electrons moving in the near axis area in the same manner as an optical lens acts on light. This property of non-homogeneous axis-symmetrical field is used to focus electron beams emanating from a photocathode on to a phosphor, producing an inverted electro-optical image of a scene that has been earlier responsible for the ejection of photoelectrons, when such a lens

117

118

An Introduction to Night Vision Technology

system is incorporated within an I.I. tube. As electrons are required to be accelerated right from their emission, the field extends right up to the photocathode surface. Technological advantages are in favour of electrostatic lens systems in I.I. tubes intended for night vision systems. That is so, as magnetic lenses tend to be bulky, heavier and with larger overall dimensions. The magnetic lenses also consume considerable power. Three types of electrostatics lenses are generally in use for the electrostatics electro-optical systems[4] (Fig. 8.7). These are: (a) (b) (c)

Aperture lens, Bipotential lens, and Unipotential lens. The aperture lens (Fig. 8.7 (a) is formed by a disc-shaped electrode with a circular aperture at a certain potential immersed in two different potentials on either side. As the field intensity varies near the aperture, it is this region which forms the lens. The bipotential lens is generally formed by two coaxial apertures, an U U1

U2

Z

(a) U1

U2

U2

U1

U2

U1

Z (b) U2

U2

U2

Z

U1

U1

U1 (c)

Figure 8.7. Scheme of electrode systems for (a) aperture lens, (b) bipotential lens, and (c) unipotential lens (U 1, U2 refer to potential values).

Image intensifier tubes

aperture and a cylinder around the same axis or by two coaxial cylinders at different potential. The potential at both sides of the lens are constant and equal to the electrodes forming the lens. The field on one side may extend right up to the photocathode when it is also referred to as an immersion objective. The optical power of bipotential lenses greatly depends on the potential ratio of the electrodes (Fig. 8.7 (b). The unipotential lens is formed by three coaxial apertures, with the two outer electrodes at a common potential (Fig. 8.7 (c). Both symmetric and asymmetric combinations are possible so that the field of the lens is symmetric in relation to the midpoint of the lens or otherwise. As in the case of bipotential lenses the optical characteristics are determined by the ratio of potential of the electrodes. In all these cases electrodes are coaxial bodies of revolution for a lens system. The subject of electron-optics has been well developed and all its correlations with the classical optics have been fully exploited. The optical equivalent path is more similar to a path through a changing refractive index medium where the refractive index n at each point gets defined by n = u with u representing the potential at the point. These potential changes take place more rapidly where a crossover or focusing action is desired. Relationships have been worked out for cardinal points, aberration characteristics, focal ratios, and the like so that the LENS REGION (FOCUSING FIELD)

n 1 u 1 r1

n 2 u 2

1

z

2 r2

a

b

Figure 8.8. Illustration of the Langrange-Helmholtz equation

119

120

An Introduction to Night Vision Technology

understanding of the subject of optics is helpful in arriving at decisions in the field of electron-optics. Analogy with optics is quite useful. For instance, one can utilize the Langrange-Helmholtz equation which is a derivation from the Abbe sine condition, and is defined in terms of the refractive indices of the media on both sides of a focusing field (Fig. 8.8). The relationship is

n 1r11 n 2r2 2

(8.3)

where the subscript 1 refers to the object space and subscript 2 to the image space, for heights, aperture angles, and refractive indices. The same relationship can be rewritten or derived in terms of potential in the form

r11 u1  r22 u 2

(8.4)

The focal lengths are also related by the formula

f1  f2

u1 n  1 u2 n2

(8.5)

A lens system consisting of two or more electron lenses can thus be defined on the optical pattern, to form effective electronoptic devices. 8.4

GENERAL CONSIDERATIONS FOR IMAGE INTENSIFIER DESIGNS A typical electron image intensifier may employ a two lens optical system[4]. The first lens besides focusing must also accelerate the photoelectrons. The field of the first lens must thus be extended to the photocathode, so as to collect and accelerate all emitted electrons, i.e., the cathode is immersed in the field originating from the potential forming the first lens. This means that the object is immersed in the field as if in a medium of refractive index n, corresponding to the under-root of the potential forming the lens in the object-space. Such a lens is also known as an immersion objective. It is essentially a bipotential lens. This may be coupled to an aperture to form a complete system. The second lens helps in the control of divergence and assists in reducing aberration characteristics (Fig. 8.9). As shown in the figure, the photocathode is immersed in the objective field. A diaphragm is provided near the crossover formed by the immersion objective. The second lens transferring the image to the screen is formed between the first and the second

Image intensifier tubes CATHODE PHOTOCATHODE IMMERSED IN OBJECTIVE CATHODE APERTURE FIELD

IMMERSION LENS

SCREEN (PHOSPHOR)

DIAPHRAGM

ANODE CONE

Figure 8.9 A typical layout of an intensifier tube

anode, and is thus a bipotential lens. The screen should be a concave surface and coincide with the surface of the best image to overcome significant aberrations, such as distortion and curvature of the image surface. The photocathode is also suitably curved for the best results for the input image. Both these aspects are taken care of in the Generation-1 tubes with input and output suitably curved fibre-optic windows. In Generation-2 tubes where microchannel plates (MCP) have been introduced the output is more conveniently coupled to plane-parallel fibre-optics windows. This necessitates an additional electrode for distortion correction before the electron beam is incident on the MCP. In all these cases while theoretical understanding and correspondence of electron-optics with the classical optic is a great help to lay down preliminary designs, the ultimate designs adopted are experimentally developed to give the best of results. Next, we can consider the design parameters for the overall intensification of distant objects that is possible utilizing I.I. tubes. Obviously, these parameters include photocathode sensitivity, luminous efficiency of the screen, and accelerating voltage, apart from the optical systems. If the light-flux (including radiation beyond the visible to which the photocathode is sensitive) is c, then it generates a current kpc where kp is the photocathode sensitivity over the entire spectral region of photocathode response. In actual practice, the sensitivity of the photocathode will be frequency dependent. The manufacturers usually give the sensitivity values separately for the white light in A/lumen and for radiation above 800 nm in A/W (The actual wavelength values may also be indicated). This current kpc gets amplified by the accelerating

121

122

An Introduction to Night Vision Technology

potential U of the I.I. tubes and transfers a power U kpc to the screen of the I.I. tube. If the luminous efficiency of the screen is , then the light flux  emitted by the screen is Ukpc. The luminous gain Gp can then be defined as

Gp 



Light flux emitted by the screen Uk p c



Light flux incident on the photocathode c 

 Uk p (8.6)

Thus, the gain is higher if the screen efficiency, accelerating potential and the photocathode sensitivity is higher. No doubt there would be limitations due to noise and dark current in the system and system components. The above is applicable if both the input object size and output image on the phosphor are of the same size. In case the I.I. tube has a magnification mi, the image would be spread over an area m i2 times the area of the image on the photocathode. Thus, we have

G p =  Uk p /mi2

(8.7)

The gain of an image intensifier is usually expressed in cd/m2/lx as a ratio of the output brightness in nits (candelas/sq m) to the input illuminance in lux (lumens/sq m). This measurement is usually done at a colour temperature of 2854°K at appropriate input light level of low order, i.e., say 20 lx. Its equivalence to the theoretical value is given by Eqn 8.7, where  the phosphor efficiency is in lumens per watt, U the applied voltage and KP the photocathode sensitivity in amperes per lumen. As the photocathode sensitivity at different wavelengths has a different value, the composition and magnitude of the light stimulus to the photocathode has to be standardized to give a consistent value for the gain. This also helps in comparing I.I. tubes from different manufacturers or in the same lot. Extending to incorporation of an I.I. tube in an instrument system and defining the total luminous gain G in terms of the object brightness B0 in a scene that is being imaged through an optical system we have G 

Brightness of the image on the screen (B s ) B s  Bo Brightness of the object (B o )

(8.8)

Referring to Fig. 8.10, we have the relationship of image heights ho, hc and hs corresponding to object, photocathode and phosphor (screen) as

h

s

mh mm h i c i o o

(8.9)

where mi is magnification due to I.I. tube as in Eqn 8.7, and m0 is the magnification due to the optical system on the photocathode

Image intensifier tubes

of the object of height h0. To investigate the total luminous gain, let us assume an area so in the object space around the axis which is imaged on to an area sc on the photocathode. The amount of light flux c that will reach this area on the photocathode will be a function of the brightness of the object Bo, the transmission factor through the atmosphere and the objective lens system , and the maximum angle of acceptance of the light cone emanating from the object area s0 , depending on the entrance pupil diameter of the optical system D and its distance from the object R or the angle 0 , i.e.,

c = B 0s 0 sin2 0 .

(8.10)

Using Abbe sine-condition we have

s 0 sin2 0 = sc sin2 c

(8.11)

Therefore, c = B 0sc  sin2 c   and if sin2 c is determined from the geometrical relationship of Fig. 8.10 the equation can be rewritten in the form

D2  D 4 f 2

0 = B 0 sc  as sin2 c =

2

D2 D 2 4f

2

(8.12)

.

This is on the assumption that the distance R of the object is very large in comparison to the focal length of the optical system and that the image is formed in the focal plane. This assumption is valid as in practice, viewing is needed for distant objects. As the image intensifier gain is Gp, (Eqn 8.6) we have

D2 s  G p .c  G p .Bosc 2 . D  4f 2 As the screen (phosphor) area corresponding to be given by

 sc m i2

Bs =

(8.13)

sc would

we have screen brightness B s given by

s

 sc m i2

= Gp 

D2 1  2  B0  2 2 D  4 f mi

giving the total gain as

(8.14)

123

124

An Introduction to Night Vision Technology OBJECTIVE (OPTICAL SYSTEM)

hO

O

PHOTOCATHODE

C

D

OCULAR SYSTEM

hS

EYE

hC f

R

INTENSIFIED IMAGE ON SCREEN (PHOSPHOR)

IMAGE THROUGH OPTICAL SYSTEM

Figure 8.10. Sketch of an instrument system with an I.I. tube

G =

Bs D2 1 = Gp 2   B0 D  4 f 2 m i2

(8.15)

This may be put in the form 2

D 1 G  0.25 G p    2  (8.16)  F  mi The numerical value in Eqn 8.16, 0.25 is really much closer to the more exactly calculated values for aperture ratios of 1:5 or slower. Its value changes more rapidly at faster F numbers. Thus at an aperture ratio of 1:1 it would be 0.20 and at an aperture ratio of 1:2 it would be 0.235. Nevertheless, the variation in the numerical value of 0.25 for different apertures is not so significant. A system with an aperture ratio of 1:1 is 25 times faster and a system with an aperture ratio of 1:2 is more than six times faster than one with an aperture ratio of 1:5 against a change 20/25 and 23.5/25 due to the more exact calculation of the numerical value of 0.25. To maximize the overall gain, the second term in this equation Gp should be as large as possible. Referring to Eqn 8.6, this would mean that the screen efficiency, accelerating potential and the photocathode sensitivity should be as high as possible. Factors relating to phosphor (screen) and photocathodes have been well discussed in the related chapters earlier. Obviously, accelerating potential and the overall design has to be such as to add minimal noise and not to overbrighten the phosphor. As the size and weight of I.I. tubes is also a major consideration, suitable power supplies (wrap-around) have also been developed with automatic brightness control. As resolution of a high gain noise limited I.I. tube is primarily limited by the finite number of photoelectrons released by the photocathode, it is advisable to design instrument systems which detect at the required distance

Image intensifier tubes

well above this limitation. This means that for an excellent tube, this has to be done primarily by having as large an aperture as permitted by various design restrictions so that more of the light flux from an object scene is concentrated on the photocathode. Thus the physical value of D 2 is important, for a practical application. The third term in the Eqn 8.16 signifies the need for a faster and faster aperture ratio. The fourth term suggests a minification of the tube magnification, i.e., the screen size to be lesser than the photocathode size. Such tubes have also been designed particularly where the overall magnification presented to the eye is around unity and larger field of view is a requirement. The final term  suggests that the transmission factor of the optical system should be as high as possible, i.e., the objective lens surfaces should be properly coated for the spectral range to which the photocathode is sensitive. Likewise, the eyepiece lenses should be coated to maximise transmission in relation to the nature of the output spectrum from the screen (phosphor)[5]. Further relevant optical considerations have been referred to in Chapter 5. As discussed therein, considerations for total field of view and overall magnification are significant. 8.5

IMAGE INTENSIFIER TUBE TYPES Further to the historical development as discussed in paragraph 8.1, and parametric details in Table 8.1, we may now discuss the types of tubes that have evolved so far. Generation-0 Image Converter Tubes These tubes referred to also as Image Converter tubes have an Ag-O-Cs photocathode with an S-1 response (Fig. 6.6). The phosphor could be a typical P-20 type. The acceleration voltage in the tubes are of the 10-15 KV order. To improve the brightness the screens of these tubes can be aluminised. This also eliminates optical feedback. 8.5.1

PHOTOCATHODE (CATHODE)

CATHODE APERTURE

ANODE CONE

AFTER ACCELERATION POTENTIAL RINGS SCREEN

Figure 8.11. Cross-sectional view of a special Generation-0 tube

125

126

An Introduction to Night Vision Technology

There were some developments in these types before the arrival of more sensitive photocathodes covering both visible and near infrared regions. There was the development of multi-slot photocathodes with higher sensitivity in the longer wavelength regions. Image converter tubes were also produced by using further after acceleration of the electrons near the screen[4] (Fig 8.11). Generation-1 Image Intensifier Tubes Generation-1 tubes started making their appearance in early sixties and had an S-20 photocathode (See Fig. 6.6) coupled through a two electrostatic lens system to a phosphor screen, usually P-20. The lens system more or less followed a similar pattern as in Generation-0 with an aperture-cone electrode combination. As the gain was not that high, it was thought expedient to cascade these tubes either internally or externally. As stated earlier, these efforts were only partially successful and resulted in cumbersome and expensive designs (Fig. 8.1 and para 8.2). A cross-sectional view of a single Generation-1 tube is shown in Fig. 8.12. Earlier versions of these tubes used S-20 photocathodes and later on stabilized to the use of S-25 photocathodes with a P-20 phosphor. As the fused fibre faceplates made their appearance, their incorporation in a Generation-1 tube made cascading relatively easier, effective and economical. Figure 8.5 shows a sectional view through a cascaded Generation-1 8.5.2

CATHODE APERTURE PHOTOCATHODE CATHODE FIBRE OPTIC PLATE IMAGE OF SCENE

GLASS WALL CYLINDER ANODE CONE

PHOSPHOR SCREEN

INTENSIFIED IMAGE

+15 kv Figure 8.12. Sectional view through a Generation-1 tube

Image intensifier tubes

tube where three tubes have been cascaded. Refer also para 8.2.2. Gains have been measured to be in excess of 30,000 and may be in a range of 50,000 to 100,000. Three-stage systems are also suitable for incorporation of automatic brightness control particularly as the system is rather sensitive to supply voltage variations and ripples. Generation-1 tube like later generations have been standardized to 18 mm and 25 mm diameters for both the photocathodes and the phosphor, thus operating at unit magnification, keeping a variety of applications in view. Tubes with 40 mm photocathodes are also in the market for specific applications. The resolution of the Generation-1 tubes is dependent on good electron-optical systems as also on the grain structure of the phosphors of the screens. As is obvious, the second and third tubes in a cascade pick up the input from the phosphor screens and may progressively degenerate the overall resolution. Excellent manufacturing and phosphor deposition or coating techniques are therefore called for. A moving object when seen may give rise to a smear. It could also cause blooming when viewing a bright object. Generation-2 Image Intensifier Tubes The second Generation tube is a combination of a singlestage I.I. tube of Generation-1 coupled internally to a micro-channel plate. The photocathode is highly improved and is of the S-25 type and has an extended red response. The micro-channel plate has been discussed above in paragraph 8.2.3. A section through a Generation-2 tube is shown in Fig. 8.13. Thus a high gain singlestage image intensifier, with a better photocathode using an electrostatic lens system impinges electrons in the input of the micro-channel plate. These electrons after intensification are shown in Fig. 8.6 are proximity focused on the phosphor screen. The electrostatic lens system has to be such as to produce a flat image at the input of the MCP. Thus the normal electrode system of a spherical cathode and a conical anode with an aperture is augmented by a distortion correction ring (a sheet cylindrical electrode) before the electrons impinge on the MCP. Impinging electrons can also generate positive ions which may travel back and reduce the life and efficacy of the photocathode. A positive ion barrier is obtained by placing the input of the MCP at a lower beam potential than the anode cone potential. A thin ion-barrier film could also be deposited on the input face of the MCP. It could trap some of the incoming electrons also and prevent the re-entry of electron that rebound from the solid edges of MCP channels on the input face. This tube has many advantages over the Generation-1 cascaded tube. It achieves the same order of lumen amplification 8.5.3

127

128

An Introduction to Night Vision Technology

INPUT FIBRE OPTIC FACEPLATE

CATHODE SHIELD - 2500 V

BODY CERAMIC

MCP INPUT - 900 V

MCP OUTPUT + 6000 V

- 2500 V MCP PHOSPOR SCREEN PHOTOCATHODE (S-25, (ERMA)

OUTPUT FIBRE-OPTIC FACEPLATE

ANODE CONE

DISTORTION CORRECTION RING

Figure 8.13. A sectional view through a Generation-2 tube

in a much smaller length and weight. As there are no phosphorphotocathode interfaces in Generation-2 tubes as in Generation-1, image smear of moving objects is avoided. Further the considerations for the graininess of the phosphor do not affect to that extent, as only one phosphor surface is involved instead of three in the case of Generation-1. However, the resolution and noise characteristics of the MCP become more important considerations in the case of a Generation-2 or more advanced tubes utilizing MCPs. A rugged wraparound voltage stabilized power supply powered by a high duty 2.7 V battery is usually used to give appropriate voltage to all the electrodes, i.e., cathode, anode cone, distortion correction ring and the screen. Circuitry also ensures flash suppression, bright source protection and automatic brightness control, to enable good image transmission. Battlefield illuminants like gun flashes, explosive, fires, etc, thus do not seriously disturb the vision. Advantage of the confinement of an illuminant to a few channels of the MCP also prevents a smear across the whole field. Successful night vision devices for low light vision have been optimised using Generation-2 tubes which are almost distortion free and have a long operational life. Dependent on the ultimate use, the system may have both a large aperture and a fast aperture-ratio to make vision possible at stipulated ranges and fields of view with high quantum efficiency photocathodes and high luminous gains available in Generation-2 tubes.

Image intensifier tubes

It is also obvious, that Generation-2 tube is an inverter like the single unit of a Generation-1 or 3 units of Generation-1 or Generation-0 and hence systems utilizing these tubes do not require any optical erector system. The inverted image on the photocathode becomes erect on the phosphor and can hence be directly viewed through an eyepiece. The very nature of focusing by an electronoptic lens inverts the image as is done by an optics-objective. Nonetheless, the Generation-2 may be named electrostatic image inverting Generation-2 image intensifier tube to emphasize its difference from the proximity focused wafer tubes. Generation-2 Wafer Tube Proximity imaging on the phosphor was tried much earlier with a view to develop the simplest types of image converters (Generation-0 type). Thus if a photocathode and a screen (phosphor) are placed parallel to each other in an uniform field inside a vacuum envelope a unit magnification image should be possible. It was shown that if the initial velocity of electrons leaving the photocathode was represented by the potential U0 and Uac was the accelerating potential for these electrons, then in a uniform electrostatic field, the diameter D of the scattering circle of the impinging electrons is given by 8.5.4

D  4l

Uo U ac

(8.17)

Where l is the distance between the photocathode and the phosphor. With practical designs this value of blur circle could not give a resolution better than 10 line-pairs per mm. Further, the backscattered light from the screen (phosphor) would illuminate the photocathode and add to the background noise. A diaphragm would also help, as the electrons moved along the lines of force of the uniform field – it could only reduce the screen area to the size of the hole in the diaphragm. A thin film of alumina of proper thickness on the phosphor could somewhat decrease the back scatter, if it permitted the forward movement of electrons and prevented the backward movement of the quanta of light. Nonetheless, the resolution was far too low to permit any devices to be built around such a system. The technological development of the micro-channel plates drastically improved on the situation as now the electrons could not only be confined within a channel, but also used to release secondary electrons improving on the gain while taking the benefit of a uniform accelerating field. The resolution would depend on the design of the MCP. It simply meant that the MCP could be now

129

130

An Introduction to Night Vision Technology

sandwiched between a photocathode and a phosphor in a vacuum envelope and an appropriate potential applied across its input and output faces. The electron defocusing as these emanated from the photocathode could be minimised by having as small a gap as possible between the photocathode and the MCP, usually of the order of 0.2 mm or less. The output from the screen could be taken out via a fibre-optics faceplate. In such a system, the resolution could be of the order of 30 line pairs per mm or better, but the photocathode image would not be inverted as in an electrostatic lens system used in Generation-0, Generation-1, and Generation-2. Incorporation of fibre-optics twister, (para 8.2.4), overcame this problem also, by providing an integrated inverting system within the vacuum envelope itself. Such tubes are now referred to as Generation2 wafer tubes as these employ the MCP as in Generation-2 I.I. inverter tubes (Fig. 8.13). Unlike the electrostatic image inverting Generation-2 image intensifiers which are available in many sizes usually with input-output faces as 50/40, 25/25 and 18/18 mm, the wafer tubes are generally confined to 18/18 mm size (Fig. 8.14). These result in highly compact and light weight designs most suitable PHOSPHOR SCREEN

CERAMIC METAL BODY

FIBRE-OPTICS INVERTER

FIBRE OPTIC FACEPLATE

MICRO-CHANNEL PLATE

PHOTOCATHODE (S-25, ERMA)

Figure 8.14. A schematic view of a Generation-2 wafer tube

Image intensifier tubes

for design of night vision goggles and night-sights for small arms. These have found a great application area in avionics also. Their freedom from distortion and uniform resolution over the entire picture area make them more suitable for biocular or binocular applications. These tubes are also referred to as double proximity focused wafer tubes because of the image transfer through the MCP which is in proximity both to the photocathode and the screen and immersed in a horizontal field. Generation-3 Image Intensifier Tubes Image intensifier tubes utilizing Generation-3 photocathodes, i.e., NEA photocathodes such as cessiated GaAs and improved MCPs are generally referred to as Generation-3 I.I. tubes. Their sensitivity to much lower light level makes them more eminently suitable for incorporation in low light level systems particularly for night vision goggles. The Gallium Arsenide (GaAs) cessiated photocathode, the photocathode of choice for Generation-3 tubes, is an excellent compromise for low dark current and good infrared detection. The photon rate is around five to seven times greater in the region 800900 nm than in the visible region say around 500 nm. However, this photocathode requires protection from bombardment by gas-ions released from the channels of the MCPs, as otherwise it would get rapidly destroyed. To avoid this effect, a thin ion-barrier film may be deposited on the entrance face of the MCP to trap gas-ions (Fig. 8.6). This film may however trap some of the incoming electrons also. A very high level of vacuum in the tube during processing would also 8.5.6

AL2O 3 ION BARRIER MGF 2 COATING (TO IMPROVE FACEPLATE OUTPUT)

FIBRE-OPTIC INVERTER

CATHODE FACEPLATE (FIBRE-OPTIC) MICRO-CHANNEL PLATE Si 3N4 COATING TO IMPROVE PHOTOCATHODE OUTPUT GaAs PHOTOCATHODE PHOSPHOR SCREEN LIGHT ABSORPTION MEDIA

Figure 8.15. A schematic view of a Generation-3 wafer tube

131

132

An Introduction to Night Vision Technology

ensure to limit the ion-feedback damage to the photocathode. MCPs may also be required to improve on their open area ratio by using manufacturing techniques which make the channel input area larger by funnel shaping their input ends. Improvements in MTF and resolution are also possible by reducing the channel diameter. The quality of MCPs and the stability of Generation-3 photocathodes become important factors to push Generation-3 I.I. tubes down for operation at 10–4 lux or a still lower value. Manufacturing accuracy and controls become very significant. Thus, the production of Generation-3 image inverter type on a regular basis is still not that frequent on a costeffective basis. Generation-3 wafer tubes, however, are now in regular use. These tubes have a close similarity to Generation-2 wafer tubes except for the type of photocathode used and may be an improved MCP (Fig. 8.15). Coatings are in use which are relevant to the type of photocathode. Fibre-optics twister to erect the image may also be incorporated as in Generation-2 [5,6]. 8.5.7

Hybrid Tubes According to some manufacturers, stability of a Generation-3 photocathode and thus the utilization of its sensitivity by a noise-limited MCP continue to be difficult. The alternative may be to prefer more robust photocathodes of the multi-alkali type coupled to low noise MCPs. The need for an ion trap (Fig. 8.6), at the input face of the MCP to restrict gas-ions from reaching a Generation-3 photocathode and deteriorating it, does reduce the total number of electrons as also the re-entry of rebound electrons from the solid edges of the MCP channels on the input face, into the micro-channel. As the total performance depends on the amplification stage MCP, photocathode and the screen improvements are more helpful by having an improved MCP with a stable advanced Generation-2 photocathode where the technology is in better control. Some manufacturers refer to these tubes as super Generation tubes. Likewise a Generation-1 inverter tube could be coupled to a Generation-2 or Generation-3 wafer tube and enable a better performance. Such tubes have been referred to as super inverters. It is also possible to gate I.I. tubes for special applications wherein the functions of an intensifier are coupled to a fast electrooptical shutter. Fast gating is used in range-gating, fast spectroscopy and in some special areas of plasma and nuclear physics. Image intensifiers, particularly the wafer types (for compactness) can be coupled to area-array charge coupled devices

Image intensifier tubes

(CCD’s). The coupling is suitably effected through a fibre-optics element, which suitably demagnifies the output size from the intensifier to that of the CCD. Many such couplings both internally and externally between different types of electron-optical tubes and image intensifiers have resulted in a number of interesting instrument systems. Some successful one’s relate to the development of low light level (night vision) television systems. Image intensifier tubes have been successfully combined to silicon self-scanning array systems resulting in suitable night vision cameras and for application in many other fields. The selfscanning array may be a charge coupled device (CCD), a charge injection device (CID) or a photodiode array (PDA). Self-scanning array-based cameras though in use independently, require to be used with image intensifiers to provide a low noise optical amplification to produce a good signal-to-noise ratio for either very low exposure applications or for operation at very low levels of light say below 0.5 lux minimum illumination, the usual limit of silicon self-scanning array TV cameras having a frame rate of the order of 1/30 to 1/25 of a second. Other important applications arise because of the ability to electronically shutter image intensifiers as fast as 1 ns or less or utilizing the higher sensitivity of the intensifiers in certain spectral regions. Suitably coupled self-scanning arrays with image intensifiers have resulted in a large number of applications be it for spectral analysis, range gating or other application of high speed optical framing cameras, military cameras, night time surveillance, and astronomy. The systems so coupled are well designed for low image distortion, linear operation, and robustness. Usually, coupling is done with 2 and 3 Generation-2 proximity I.I. tubes. It is also possible to operate with two or three micro-channel plates in face to face contact to achieve high electron gains in an I.I. tube. While the electron gain could be more than double by such means, the resolution would have a tendency to fall and almost get halved. Such systems, i.e., image intensifiers coupled to silicon self-scanning array systems or CCD’s either optically or preferably through fibre-optics have also been used for active imaging. A narrow beam CW laser raster-scans across the object scene and its reflections are displayed by the system on a video monitor for direct viewing. The system can be operated in atmosphere, under water or in space. The advantage lies in scanning large fields of view over very short periods of time. A more interesting method employs a laser pulse of only a few nanoseconds synchronized with a gated I.I. based CCD so that only reflected pulses corresponding

133

134

An Introduction to Night Vision Technology

to a certain distance alone are received. Thus, the imagery is both free of noise contributed by the interviewing medium and well focused for the stipulated time or distance. This enables the exact range also to be known. The method is referred to as range-gating. Another interesting application using such systems could be day-cum-night cameras. This could be done by introducing autoiris camera lens to control the effective aperture of the optical system and by controlling the MCP gain. Thirteen orders of luminance- magnitude are known to have been automatically covered by such cameras. Such systems have also been used in spectroscopy to measure optical radiators in linear patterns. Conventional methods of measurement, i.e., by photographic films, single channel photomultiplier tubes or by TV camera tube has given way to image-intensified charge couple devices. Spectra are acquired up to 1000 times faster and/or with better signal-to-noise ratio during a given measuring period. Applications are for Raman spectroscopy, multiple input spectroscopy and small angle light scattering. 8.6

PERFORMANCE OF IMAGE INTENSIFIER TUBES A number of manufacturers internationally produce I.I. tubes mostly for incorporation in night vision devices for different applications that might have been designed or produced by the same set of manufacturers or others. The acceptance of these devices is done through standard specifications which might have been laid for each type of instrument and the tube. Besides the optical and electro-optical performance, the I.I. tubes have to be environmentally stable, withstand extremes of climates, be of minimal size and weight and be cost-effective for the application in view. The important parameters of optical and electro-optical significance are (i) signalto-noise ratio (ii) modulation transfer function (resolution and contrast), (iii) output brightness and its uniformity, (iv) automatic brightness control, and (v) its life. Factors like image shift, image alignment, and equivalent background illumination are also of concern in the tube as a whole. Besides these and simulator tests, evaluation has also to be done for independent testing of photocathode and phosphor sensitivity and verification of electric stability[7,8]. 8.6.1

Signal-to-Noise Ratio At very low levels of illumination, the statistical variation in the photon stream becomes more dominant and this results in quantum noise for an elementary image area depending on the number of photons received by it. When such a stream is incident

Image intensifier tubes

on a photocathode, the resolution characteristics of the tube are limited primarily by the number of photoelectrons that have been emitted as also their statistical distribution. Because of this, the intensified image of a discrete object may not be recognizable as it could be broken into an assortment of scintillations for that order or resolution which is limited by the statistical considerations of the photoelectrons received on the phosphor, even when integration over the storage time of the eye in relation to phosphor characteristics may be a little helpful. In practice this limit of resolution would be further limited because of the photocathode quantum detection efficiency, photocathode noise current, statistical distribution of photoelectrons after multiplication in an MCP or the tube noise factor – an overall measure. This low light level resolution is proportional to the signal-to-noise ratio which can be defined as the ratio of the ‘dc signal to the r ms value’ in the output beam. For precise comparisons and evaluation, these measurements will have to be done at a specified very low light level input which may correspond to a starlight or overcast sky over an area which may be of the order of a pinhole. As the I.I. tubes may be active even when no light is incident, these measurements require to be modified suitably. Thus the signal-to-noise ratio (S/N) may be defined as

S o  Sb S  N (N o2  N b2 )1/ 2

(8.18)

where So =

dc signal output when the tube is illuminated at the specified level of illumination

No =

rms noise output at the same specified level of illumination

Sb =

dc signal when there is no input light on the I.I. tube, i.e., the background signal

Nb =

rms noise when there is no input light on the I.I. tube, i.e., the background noise

A constant of proportionality K is also introduced which is dependent on the phosphor decay characteristics and involves a correction factor to obtain a signal-to-noise ratio over an equivalent bandwidth of 10 Hz independent of the frequency response of the assembly. The equation is thus rewritten as S /N 

1 So  Sb K N2 N2 o b





2

(8.19)

135

136

An Introduction to Night Vision Technology

Measurements are done with special test equipment which utilize low dark current photomultiplier tubes and are able to measure the dc and r ms values over an electronic bandwidth of 10 Hz. Thus, S/N ratios of the order of 3:1 or better may be achieved when illuminating an area of an order of 0.2 mm dia on a photocathode at illumination levels of the order 1.2  10–4 footcandles. Lower sensitivity of photocathodes, increasing MCP voltage, a proper open ratio of the MCPs ion feedback effects, and a poor detection efficiency of the phosphor and similar defects can all lead to a poor S/N ratio[8]. 8.6.2

Consideration of Modulation Transfer Function (MTF) Assuming an I.I. tube to be a linear system, the total MTF of an I.I. tube would be a multiplication of the MTF values of its components, i.e., the photocathode, electron-optics, MCP, the screen (phosphor) and the fibre faceplates[3]. Thus, the overall MTF of an I.I. tube can be written as MTF (overall of I.I. tube) = MTF (fibre-optics input faceplate) MTF (photocathode)  MTF (electrostatic lens system)  MTF (MCP)  MTF (screen)  MTF (Fibre-optics output faceplate) (8.20) MTF deterioration due to fibre faceplates and the photocathode is relatively insignificant as the centre-to-centre distance in these fibres is of the order of 5 m. This may not be so true for fibre-optics image inverters where the centre-to-centre distance may be to an order of 10 m. The MTF of the electrostatic lens system can be considerably improved, as already stated by curving the input and output surfaces which is very practicable with fibre-optics faceplates. Where an output from the electrostatic inverter tube is desired to be focused onto the plane face of the input to a MCP, freedom from distortion is obtained by introducing field-flattener electrodes. Thus, while electrostatic system is responsible for a little reduction in the overall MTF, it still is not the limiting parameter. Keeping this in view one can say that the MTF of the Generation-1 single stage tube is limited by the MTF of the screen (phosphor). As the electrostatic lens system and the photocathode have almost an MTF value of unity, it is obvious that a good phosphor efficiency is an essential requirement. It is well known that the detection efficiency of these screens can vary from 50 to 90 per cent depending on the manufacturing process. Hence,

Image intensifier tubes

we have MTF for Generation-1 single tube mainly limited by MTF for the screen, i.e., MTF (Generation-1, single tube) limited by MTF (Screen)

(8.21)

and in the three-stage version we have MTF (Generation-1, three-stage) limited by MTF (Screen)3

(8.22)

In Generation-2 inverter tubes in addition to a similar limitation as on a Generation-1, we have more restrictive limitations due to MCP. The limitation due to MCP will be both due to its physical configuration as also dependent on its gain-parameters (para 8.2.3). Thus, we have MTF (Inverter tube with MCP) limited by MTF (MCP)  MTF (Screen)

(8.23)

This would apply to Generation-3 and Generation-2 proximity tubes, except that MTF (electrostatic lens system) would not be relevant in this case. Standard methods of measurement are in use at specified low light levels may be of the order of cloudy moonless nights, usually at the centre of the image screen. The normalisation may be with respect to a spatial frequency of a low order say 0.2 line pairs per mm. The specifications would then lay down the acceptance values at increasing lp/mm, values with obviously decreasing percentage values. This it may state acceptance values of 25, 60 and 90 per cent at 15 lp/mm, 7.5 lp/mm and 2.5 lp/mm. Notwithstanding these MTF measurements for the tube as a whole, criteria need to be laid down for centre and peripheral resolution also at low light levels, where the variation should be minimal and resolution in excess of 30 lp/mm. Suitable tests are also designed to check that the resolution does not fall off seriously at higher light levels and obliterate vision of low light level objects in the neighbourhood of a relatively intense object. This could be also an indirect measure of veiling glare. 8.6.3

Luminous Gain & E.B.I Exact procedures are laid down to measure luminous gain at various luminous input levels and also evaluate the equivalent background illumination (EBI) at room temperatures and may be at stipulated high and low temperatures to satisfy military requirements. It may be noted that EBI is an optical measure related to the minimal brightness level of the photocathode

137

138

An Introduction to Night Vision Technology

corresponding to a light level around a decade or so below an overcast sky in relation to when there is no incident light on the I.I. tube. In a way it is a measure of the optical dark current. Thus, if Ip is the dark current of a sensitive photomultiplier, Io the current in the photomultiplier through the intensifier when there is no light on the photocathode, IB the photomultiplier current due to the brightness of the intensifier at a very low level corresponding to a decade or so less than the overcast sky, say 210–11 lumens/cm2 we have

EBI =

Io I p I B Io



= 2 10-11 lumens/cm2



(8.24)

Obviously, IB should be significantly greater than Io so that the tube performs reasonably well at higher levels of illumination. The value of the fraction Io–Ip/IB–Io has been put at one or less than one in some specifications [8]. 8.6.4

Other Parameters There are many more parameters that are required to be tested particularly in relation to the prolonged use of an I.I. tube. Thus, automatic brightness control and freedom from damage on exposure to brighter sources, electrical stability, surface properties of the photocathode and the phosphor, and the like are required to be appropriately tested. Likewise, tests have been devised for photocathode stability, as also the stability of the tube over a wide temperature range. Environmental stability has also to be ensured to a high degree of reliability not unlike other equipment of a sophisticated nature used by the military. 8.6.5

A Note on Production of Image Intensifier Tubes In view of all the requirements discussed above, it is apparent that the production of I.I. tubes has to be carried out under a strict control both while selecting suitable materials for component making and in assembly. Though it is not proposed to go into the details of manufacture it is obvious that all the considerations that are applicable to the high order vacuum tubes are of great relevance to these tubes also, apart from considerations for deposition of the photocathodes in vacuum, application of phosphor and integration of input and output fibre-optics windows as also the MCPs. Usually, the fibre-optics components and the MCPs are not made in-house but purchased from other source or a subsidiary unit. The design and assembly has to ensure that the photocathodes do not get poisoned and further avoid burnouts in any of the sensitive surfaces. The details of the manufacture are

Figure 8.16. A view of module assembly room

Image intensifier tubes 139

140

An Introduction to Night Vision Technology

Figure 8.17. Cathode processing station

considered to be trade secrets even when these methods differ from one production centre to another. The other area of specialisation in production was the wrap-around power supply which took sometime to evolve before becoming a routine fitment. Likewise, the assembly also has to be implemented under very strict environmental conditions. Thus, one manufacturer reports that the modular assembly is carried out in class 100 laminar flow clean air tables housed in class 10,000 environment (Fig. 8.16). The multialkali photocathode is processed under ultra high vacuum of the

Image intensifier tubes

Figure 8.18.

Assembly of I.I. tubes: 18 mm and 25 mm bare tube modules, high voltage power supply units and finished goods.

order of 10–9 mm of Hg. The vacuum system is an all stainless steel chamber with vacuum manipulators employing cryo pumps (Fig. 8.17). As ultra high vacuum techniques and photocathode deposition techniques are more or less well established, the chain of production of I.I. tubes and their testing does give a high rate of acceptance unlike what happens with quantum detectors particularly in the linear or matrix form for night vision in the thermal region of 8-12 m. The I.I. systems thus continue to be cost-effective for a large number of applications. Finally (Fig. 8.18) shows the relatively simpler subassembly schemes of 25 mm and 18 mm I.I. tubes. The upper row in the photograph shows the subassemblies that go to form the complete 25 mm I.I. tubes, while the lower row shows a similar layout that form the 18 mm I.I. tube. The first column shows the wrap-around power assemblies for 25 mm and 18 mm I.I. tubes, respectively. REFERENCES 1. 2.

Biberman, L.M., & Nudelman, S., (Eds). Photoelectronic Imaging Devices. Vol. 1 & 2. (Plenum Press, 1971). Kapany, N.S., Fiber Optics: Principles and Applications. (Academic Press).

141

142

An Introduction to Night Vision Technology

3. 4. 5.

6.

7.

8.

Kingslake, R. & Thompson, J. B. (Eds). Applied Optics & Optical Engineering. Vol. 6, Chap. 10. (Academic Press. 1980). Zhigarev, A. Electron Optics and Electron-beam Devices. (Moscow: MIR Publishers, 1975). Csorba, P. I. Current Status and Performance Characteristic of Night Vision Aids, in Opto-Electronic Imaging. (New Delhi: Tata McGraw Hill Publishing Co., Ltd., 1987). Girad, P.; Beauvais, Y. & Groot. P.D. Night Vision with Generation-3 Image Intensifiers, in Opto-Electronic Imaging. (New Delhi: Tata McGraw Hill Publishing Co. Ltd., 1987). Cochrane, J.A. & Guest, L.K.V. Image Intensifier Design Technologies, in Opto-Electronic Imaging. (New Delhi: Tata McGraw Hill Publishing Co. Ltd., 1987). Image Intensifier Assembly, 25 mm. Micro-channel Inverter, MILI-49040E. (Military Specification, 29.5.92.).

CHAPTER 9 NIGHT VISION INSTRUMENTATION 9.1

INTRODUCTION The image intensifier (I.I.) based instrument systems developed so far have been of significant use in night time observation and navigation, primarily on land and from helicopters. The need for night time use to direct fire on enemy targets by the infantry, artillery and the armoured corps has resulted in a series of instruments for each specific application. It is therefore obvious that the instruments systems are likely to have optical characteristics like the field of view, magnification etc., similar to those in use during daylight for observation, navigation and fire control. Reticles would also be required to be introduced for proper laying and engagement and thus match the weapon capabilities as accurately as possible. The methods of mounting on or in the weapon system is also of great concern. Besides, like all other types of military instruments these instruments have to withstand climatic and environmental tests as may be laid down for instruments in the daylight category for a given weapon system. These requirements would be both for use and in storage. The criteria for acceptance of I.I. tubes laid down in military specifications, as also for the acceptance of I.I. based instrument systems, include these aspects in detail. Further, as we are aware of the limitations of the human eye, environment and night conditions, technological aspects get mainly augmented by optical considerations and the I.I. tubes. Image intensifier tubes in turn are dependant on photocathodes, electron amplification and phosphors. The instrument system as a whole is therefore an integrated multiple of all the above factors. Nonetheless, the success of an instrument would depend on overall considerations intended for the satisfaction of a user. Apart from field of view, magnification and the mechanical limitations that it may have to satisfy, it is obvious that the user would be interested

144

An Introduction to Night Vision Technology

in the distance that such a system can see during the night, i.e., the night range. This is an important parameter of a system which cannot be predicted by any single subcomponent and will be also dependent on the night time conditions. Theoretical and experimental prediction about the night range is therefore an important parametric requirement. 9.2

RANGE EQUATION Various paradigms have been developed from time to time to arrive at a possible range value during night time. One such paradigm[1] has been explored here more to illustrate the factors on which range is dependent and to indicate the possibilities for optimisation. This theoretical approach gives a reasonable basis but it is still necessary to evaluate a given system under standard night time conditions so that the range in the field can be more or less estimated to a reasonable degree of accuracy. It would still be an estimate even when tried out in practice in the field as the field conditions are not likely to remain standard all through the measurements. If we take an object dimension of Z m at a night range of R m and assume that N line-pairs at spatial frequency Ak in line-pairs per mm are required to detect it at the photocathode we have,

Z N / Ak  .......(in meters ) R F

(9.1)

where F is the focal length of the objective in mm. This relationship though geometrically true requires to be investigated further for the practical value that R can attain for a given night vision system. If we now concentrate on a detail of area a = Z 2 in the image at the photocathode as the minimum area of detection and assume bar chart as an object in both the x and y planes of the image where N/A k is the minimum resolution in each direction, i.e., x and y, we have in a rotationally symmetric optical system: a=N/Ak .N/Ak .10 6 m

or a= (N/Ak )2.10 6 m

(9.2)

Assuming a photon flux of n1 photons from the object per sq. m on the detail then over an integration time of t seconds

Night vision instrumentation

for the system as a whole, we have the number of photons per integration time incident in this area as

n1at

(9.3)

and further assuming a photon flux of n2 photons on this area, for the same integration time from the background, we have the total number of photons incident per integration time as

n 2at

(9.4)

From the above the signal, S can be defined as

S  n1at  n 2at

(9.5)

and the noise N as the quadrature sum of photons in the detail and the background, i.e.,

N  n1at  n 2at

(9.6)

due to photon fluctuations both in the detail and background. Thus, the signal-to-noise ratio (S/N) is given by (n  n 2 )2.at (S / N )2  1 n1  n 2 2 = C n1  n 2  at

= 2C 2n at

(9.7)

Where C the contrast is defined by n1  n 2 / n1  n 2 , and n is the average number of photons of both the signal and the background photon rate. The deVries-Ross law then states that the detail can be resolved if S/N is greater than a value p which is a constant dependant on the type of the scene and the state of eye and thus a factor related to perceptibility, i.e., 2C 2n at  p 2 or in the limiting case 2C 2n at  p 2

(9.8)

For a bar pattern this value can be taken as 1.1 according to some authors. Now, if the mean illuminance at the photocathode is c lux and the photocathode sensitivity is Kp in A/lumen, the current generated then is c.Kp microamperes. The number of photoelectrons emitted therefore is

145

146

An Introduction to Night Vision Technology

c .K p

electrons/s/m 2

e.

(9.9)

where e is the electron-charge in C having a value of e =1.6010–13 C. Further if the noise power factor of an I.I. tube is defined by  Signal to noise ratio of the photoelectrons  f   Signal to noise ratio of the output scintillations 

2

(9.10)

the effective number of photons available for detection is

c . K p e. f

. per s/m 2

(9.11)

This factor f would incorporate the efficacy of the MCP and the phosphor screen apart from other factors such as that be contributed by the electron-optics of the tube. Substituting this value for n in Eqn 9.8, we have





2C 2 c  K p at e. f

= p2

and further substituting for a from Eqn 9.2, we have f 2  AK /N  C K p t

c  A 

where A 

2

(9.12)

ep 2  6 10 is a constant. 2

Following Eqn 8.16, c has a relationship with object luminance o. The object luminance is given by

 o  0.25  o D/F    to a first approximation 2

0.25  0 (F number )2. .

(9.13)

where D,F, and  have the same meaning as in para 8.4 of chapter 8, i.e., diameter of the objective D, focal length of the objective F, and  the transmission factor through the atmosphere and the objective lens system.  further adds a factor indicating the reflectivity of the object scene. Combining Eqns 9.12 and 9.13, we have

Night vision instrumentation

f A 2 A /N 2  0.25  0 F number 2   C K p t k Substituting for  Ak /N  A.

2

from Eqn 9.1, we have

f (R / Zf )2  0.25 o (F . number )2.  .  C . K p .t 2

i.e.,





R 2  A *Z 2  0 F number  F 2 C 2 k p /f   t 2

(9.14)

where A* = 0.25/A and basically taken into account the perceptibility factor p, the electron charge e and other numerical values, taking care that the resultant value of R is in metres. Before one makes use of this equation, it would be better to interpret C the quantum contrast in terms of the modulation transfer function of the night vision system as a whole and of its constituents. The quantum contrast can be directly interpreted in terms of normally defined contrast

i .e.,

I max  I min I max  I min

Thus, while viewing, a line-pair Imax would signify the average brightness in the white bar and Imin the average brightness of the dark bar. Restricting now to a frequency Ak line-pairs/ mm, we could state that the object contrast Co has been modified by the modulation transfer function (MTF), M of the total electrooptical night vision system. While this may not be quite true for square wave response, as M refers to sine wave response only, it has been shown that for frequencies of normal interest, i.e., higher than 2.5 line-pairs/mm the higher harmonics in the expansion of a square wave in terms of a summation of sine waves do not have greater significance. Further, as a night vision system can be taken to be a linear optical system, the total MTF, i.e., M can be obtained from the MTF value for the individual subunits of such a system. Usually, as a night vision system consists of three major subsystems in a cascading order, it can be stated that M= M o. M i. M e

where, Mo is the MTF of the objective subsystem

(9.15)

147

148

An Introduction to Night Vision Technology

Mi is the MTF of the I.I. tube

Me is the MTF of the eyepiece

All at a given frequency Ak.

Equation 9.14 can now be rewritten in the following form:



R 2  E .( Z 2 . o . . ) (F number )2 .F 2 .M o2 .M e2

 K

2 p .M i / f

.t

(9.16)

where constant E now additionally takes into account the constant of proportionality between the contrast values and the MTF. The above equation is illustrative of the fact that the range achieved in meters depends on the following factors: (i)

Constant E which takes into account the perceptibility factor, electron charge and constants of numerical conversion from sq. mm. to sq metres.

(ii)

Factor: Z 2.o .  .  This factor could be considered as the object scene factor, as it refers to Z the minimum detectable size of the object scene,  the reflectivity of this scene and  the transmission factor through the atmosphere and the objective lens system. As o. and  are natural factors, one can only argue about the Z value that may be necessary in practice for detection of given objects, assuming that the transmission factor through the optics has already been optimised.

(iii)

Factor:

F number  F 2

2

M o2 M e2



This could be considered as the optical factor. Herein we find that range is directly proportional to F number and the focal length value. Thus range will be higher if the F-number is faster and focal length larger. This could also be interpreted as that the diameter of the objective should be larger. However, as considerations of systems design particularly their requirements of compactness as also for field of view limit the diameters to which one can go practically, it is obvious that for a given diameter, the system should be as fast as possible, compatible with the desired field of view requirements and the values desired for the MTF. The optical factor also indicates that for the spatial frequencies of interest the values of MTF for the objective and the eyepiece should be as close to unity as possible.

Night vision instrumentation

(iv)





2 Factor K p M i / f t

This factor refers to the image intensifier function. The higher the sensitivity of the photocathode, i.e., Kp, the better will be range achieved provided the noise factor f does not increase proportionately. The designers therefore try to balance these two factors to achieve the best response that is possible. Low noise MCPs have been used to improve on the Generation-2 tubes to enable such types to compete with Generation-3 tubes (para 8.5.7). As in the case of MTF values for the objective and eyepiece, here also the attempt has to be made to improve on the MTF value of an I.I. tube. Integration time t however is decided by the phosphor time response which in no case should be more than that of the eye to detect movement. Many authors have worked out similar relationships as in Eqn 9.16, essentially trying to calculate range for different low light level objects with varying atmosphere, optical and intensifier characteristics. Some of these are also amenable to computer programming. Nomographs have also been evolved. One would now like to invite attention to our assumption as it was stated earlier that an object of dimension Z m is detectable at the photocathode at a spatial frequency Ak with N line-pairs. The value that N should have is not clear from the above equation and rightly so, as this value is dependent on our ability to detect, recognize and identify the object. Reference to Johnson criteria in para 2.4 is therefore now relevant. As stated therein, the minimum line-pairs for detection, orientation, recognition and identification are 1.0, 1.4, 4.0 and 6.4 line-pairs respectively with tolerances as indicated therein. Thus N has a value dependent on the task that has to be performed. Infantry, artillery, armoured corps and other wings of the Armed Forces define their own target and target sizes in relation to the range and accuracy of their weaponry. Hence the value of N, i.e., the number of line-pairs at a particular spatial frequency should be such as to be able to detect, recognize or identify in accordance with Johnson’s criteria for the target-size in question. For instance tank detection is usually decided around the turret-size and assuming an accurate engagement at around 2 km range, one should be able to detect less than 2 metres at a still larger range. This information can be correlated with the spatial

149

150

An Introduction to Night Vision Technology

frequency and the number of lines that may achieve such an objective and hence the minimal parameters of the night vision system. Intensive work has been done at many places particularly in the USA to arrive at the possibility of detection and recognition as a function of system resolution across critical target dimension. Thus, it has been worked out by one group of researchers that a 60 per cent probability for recognition exists for a vehicle if three cycles of spatial information are resolved in the height of the vehicle. Similarly, a similar value of possibility may be thought of for detecting a human being at around two cycles of spatial information (Fig. 2.1). 9.3

EXPERIMENTAL LAB TESTING FOR RANGE EVALUATION While one can project the optimal ranges using equation at 9.16 or approaches similar to it, the values obtained can only be a good guide for a design effort. It would still be necessary to experimentally lab-test such designs with artificially created night vision scenes before operating these in the field. Workers in this specialization have therefore evolved experimental methods for the purpose. Each such research group or a quality control agency is likely to work out the exact conditions for system evaluation based on their own experiences. One such group in India preferred to utilize a hall of around 300 feet in length and 24 feet in width properly light-sealed for this purpose[3]. Their approach involved (i) study of illumination characteristics of artificial sources and their matching to natural levels of illumination during night, (ii) making of models/targets with different reflectivities and contrasts which could be placed at one end of the hall, and (iii) using the other end of the hall for installing instruments to observe the models/targets. The details were worked out in the following manner: (a)

Study of illumination characteristics of artificial sources and their matching The spectral distribution and radiant power of night sky was studied and the distribution was standardised as shown in Fig. 4.2 both for moonlight and starlight. Thereafter, efforts were made to combine suitably attenuated low power tungsten sources with appropriate filters such that the resultant transmission had both the spectral and intensity distribution corresponding to the night sky as shown in Fig. 4.2. A cluster of three such lamps and filters with suitable apertures was selected to produce reliable levels of illumination ranging from

RADIANCE WATTS/sq cm/STERADIAN/m

10–10

10–9

10–8



0.4

 0.6

0.8

1.0



1.2 1.4 WAVELENGTH (m)

1.6

1.8

2.0

LOW POWER TUNGSTEN LAMP THROUGH SELECTED FILTER C – 

Figure 9.1. Three lamp cluster as an equivalent to starlight



LOW POWER TUNGSTEN LAMP THROUGH SELECTED FILTER B – 

LOW POWER TUNGSTEN LAMP THROUGH SELECTED FILTER A – 

STARLIGHT





SUM OF THREE LAMP CLUSTER

Night vision instrumentation 151

RADIANCE WATTS/Sq cm/STERRADIAN/m 10-10

10-9

10-8

0.4

'

'

'

0.8

1.0 1.2 1.4 WAVELENGTH (m)

1.6

1.8

2.0

MOONLIGHT

Figure 9.2. Three-lamp cluster as an equivalent to moonlight

0.6

LOW POWER TUNGSTEN LAMP THROUGH SELECTED FILTER C' - '

LOW POWER TUNGSTEN LAMP THROUGH SELECTED FILTER B' -  '

LOW POWER TUNGSTEN LAMP THROUGH SELECTED FILTER A' - '

'

'

'

SUM OF THREE LAMPS

152 An Introduction to Night Vision Technology

Night vision instrumentation

clear moonlit night to overcast starlit night. Twenty-four such lamps were housed in an illuminator of a suitable design painted inside with dull, white diffusing paint to give uniformity of illumination over a given field of view. Ten such illuminators were placed strategically at a suitable height along the length and width of the hall at predetermined places such that there could be uniform illumination in the scene area as also its foreground. A selection of eight levels of illumination was possible to correspond from overcast starlight to full moonlight. While the first four levels corresponded to starlight with varying amount of cloud-cover, the next four levels represented increasing moonlight levels with the phases of the moon. Figures 9.1 and 9.2 respectively depict the illumination levels as obtained in this hall at the target end in comparison with the standard illumination levels for starlight and moonlight. (b)

Making of models/targets with different reflectivities and contrasts for the scene-area The approach here was two-fold. One was to utilize standard established large size resolution charts like those based on USAF 1951 charts with decreasing spatial frequencies at three levels of contrast, i.e., high, medium, and low, to work out the limits of the systems to be tested. The other was to prepare models of likely objects, such as vehicles, tanks etc, for placement in the scene area for observation from the other end of the hall. The models subtended the same angle at the observation point, i.e., at 300 feet as an actual object would have subtended at a pre-thought of range in kilometres. The models thus were scaled down from the originals both in respect of size and contrast as also reflectivity. Results could be thus obtained which would be closer to those in the field. These models and test charts were placed at a suitable height by putting them on a stage where the foreground on the stage could also be suitably illuminated and/or varied in its reflectance. The stage area was almost as wide as the hall and with a length equal to its width. This ensured a reasonable field as well as an adequate foreground for the models and test charts when tested from the other end.

(c)

Installation of instruments Suitable pedestals were made at the observation end of the hall for convenience of observation of the scene area at the other end. Naturally, the heights of the pedestal and the stage

153

154

An Introduction to Night Vision Technology

have to be appropriately matched. The night vision instrument testing in this manner proved quite effective in comparing various designs for various applications. The data compiled with the use of eight levels of illumination from clear moonlight to an overcast sky provided the field behaviour of the instruments to a sufficiently accurate degree to satisfy user requirements. Another approach has been to prepare a collimator type test equipment[4]. Thus, an integrated test equipment designed by one of the agencies is a composite unit having facility of producing the required low light level collimated beam for resolution and gain measurements and other tests. It uses a calibrated lamp, an iris diaphragm, a Lambertian integrator and a set of neutral density filters for producing low light levels to illuminate the reticles of a collimator which can be brought into the focal plane one by one using a rotating turret. The reticles introduced may be USAF 1951 pattern, so as to check on resolution of the night vision device at low light levels varying from cloudy starlight to full moonlight at different contrast levels or a uniformly illuminated plane parallel glass window which becomes a source for gain measurement by the system as a whole. A suitable photometer can be used to measure the light levels incident on the objective lens of the night vision device. With various accessories, such a system can be used to check the overall optical and electro-optical parameters of the night vision system or I.I. tubes, independently. The approach thus gives a good result on the optical and intensifier factors referred to in Eqn 9.16. It can be utilized to test a series in production or to finalize the attributes of an acceptable design. The advantage is that it dispenses with the need of a long hall and the equipment can be used both for quality control and production. It however, cannot lead to direct prediction of ranges possible as in the former approach where models are in use, subtending the same angle at the observation end as an original would have subtended at a given range. 9.4

FIELD TESTING Notwithstanding the lab-testing or range-prediction as referred to in earlier paragraphs, it may still be necessary to field test every new instrument both from the point of view of night observation and night fighting capability after their successful laboratory tests. While for night observation one can select any desired area and conduct experiments both in moonlight and starlight during different periods, the same is not true for checking

Night vision instrumentation

their night fighting capabilities. For such testing it would be necessary to actually fire a weapon system to arrive at a correct decision. Obviously, such tests will have to be done in proper established military ranges otherwise utilized for daylight engagements also. Once a new instrument is approved on the field, further serial production can be very well controlled by laying down the lab standards and by checking the overall performance by lab testing. 9.5

INSTRUMENT TYPES Figure 9.3 shows a photograph of a Generation-0 night vision system which was adopted in early sixties for the armoured corps. As indicated in the photograph, the complete set consisted of a driver’s night sight with infrared headlights cutting out the visible as also the gunner’s night sight which operated in tandem with the larger infrared searchlight cutting out the visible and thus permitting larger ranges to be achieved with the sights. Obviously, such a system could be detected with infrared sensors. Further, the system was rather cumbersome as a large searchlight had to be barrel or turret mounted involving many complications in the mechanical design so as to confine both the sights and the searchlight to the GUNNER'S NIGHT SIGHT

POWER SUPPLY UNIT

COMMANDER'S NIGHT SIGHT SEARCHLIGHT (GUNNER & COMMANDER)

DRIVER'S NIGHT SIGHT

HEADLIGHT (DRIVER)

Figure 9.3. Image converter based active night vision devices

155

156

An Introduction to Night Vision Technology

Figure 9.4. Hand-held binocular (image-intensifier based)

same area of vision. Obviously, for driving the system, one has to have a unit magnification and a significant field of view, while for engagement of a target the Gunner’s sight should have a compatible magnification and correct illumination of the scene. Consequent on the development of alkali and multi-alkali photocathodes and Generation-1 and Generation-2 series of instruments, Generation-0 series is now obsolete for military purposes, though these can still be used for perimeter search or surveillance in security zones. Figure 9.4 shows a hand-held night vision binocular of the Generation-2[5]. Such binoculars utilize advanced I.I. tubes matched for their sensitivity and noise factor or tubes of Generation-3. Generally, these binoculars have only one objective channel while the viewing is through two oculars, more for comfort of vision than for detailed depth appreciation. For some applications, the phosphor screen can also be viewed through a carefully designed ocular system allowing the scene to be seen with both the eyes through a single magnifier type of optical component. Such systems referred to as biocular systems have a distinct advantage, as the positioning of the eyes is not critical. Figure 9.5 shows an I.I. based night vision observation device integrated with a laser rangefinder and a goniometer[5]. Such an observation device has a large aperture at a fast f-number so that the optical factor in Eqn 9.16 is maximized in addition to the image intensification factor which in any case should be as high as possible. This approach permits a maximum night range that can be viewed subject only to the object scene factor. The object scene is also improved by appropriate coatings of the optical elements so

Night vision instrumentation LASER RANGEFINDER

IMAGE INTENSIFIER BASED NIGHT OBSERVATION DEVICE

GONIOMETER

Figure 9.5. Image intensifier-based night observation devices integrated with laser range finder and goniometer.

that the transmission factor through the optics is also as near unity as possible. As the night vision capability is high, this capability can be utilized for fire control purposes by say an artillery unit if both range measurement and direction of the target are simultaneously measured. Thus, while the laser rangefinder mounted in tandem is utilized to measure the range accurately utilizing laser pulses, the goniometer gives a measurement of the precise bearing (and may be elevation too). Obviously, the two units, i.e., the night observation device and the laser rangefinder, have to be appropriately synchronized and then appropriately mounted on the goniometer to aim at a given target. Figure 9.6 shows a low light level television system [5]. Though low light level television systems have not been discussed in this monograph in depth, these have found limited use in applications. It is particularly useful where the same scene is required to be viewed at different places simultaneously, as for instance a

157

158

An Introduction to Night Vision Technology

Figure 9.6. Low light level television system

commander and a gunner in an armoured fighting vehicle. In such systems the I.I. tube may be coupled to a vidicon or orthicon type of tube either internally or externally. Presently, however, it is usual to couple image intensifiers to silicon self scanning array systems like the charge MONITOR (OPERATOR)

MONITOR (COMMANDER)

Figure 9.7. Thermal imager

CAMERA

NIGHT

(ROLE OF NIGHT VISION)

COMMANDER'S EPISCOPE (DAY LIGHT)

LOADER'S EPISCOPE (DAY LIGHT)

Figure 9.8. Fire control system (role of night vision)

DRIVER'S NIGHT SIGHT

DAY CUM SIGHT

COMMANDER'S

GUNNER'S ARTICULATED SIGHT (DAY LIGHT)

GUNNER'S DAY CUM NIGHT SIGHT

Night vision instrumentation 159

160

An Introduction to Night Vision Technology

couple devices to obtain optimum utility for a series of applications (para 8.5.7). Night vision based on image intensification has a strong competition from thermal imaging. Thermal imaging for night vision utilizes the atmospheric window 8 to 12 m in the far infrared. Self-radiation from the objects in this region is collected through a fast infrared optical system and concentrated on a quantum detector or detectors in some arrangement, i.e., array, columns or matrix sensitive to that region. Dependent on the design considerations of the system, the instrument system may employ scanners, cooling arrangement for the detectors and coupling to a video system for display. The intricacies of the design as also the cost of detectors along with their selective electronics and other opto-mechanical and opto-electronic considerations besides cooling systems, scanners and the like make the system relatively very expensive. Nonetheless, the alternative is well utilized for specific applications where cost considerations are not that important vis-a-vis the strategic requirements for long range viewing. Night vision based on thermal imaging has not been dealt within this volume. Figure 9.7 shows one such system developed by the Defence R&D Organisation[5]. As may be inferred, the refractive materials for optics for the infrared region have also to be such as to transmit in the region 8 m to 12 m and thus are special both in their nature and in their working.

Figure 9.9. Weapon-mounted night vision device

Night vision instrumentation

It is natural that with the development of the night vision technologies, weapon systems have also been adopted to incorporate these features to extend their roles for night fighting. Thus, an armoured vehicle should have night vision capabilities for its movement, observation and target engagement. In other words, the driver should have a night sight, a gunner too, but one with an additional capability to guide engagement to a target while the commander has a system for night time observation. Figure 9.8 shows a sketch of an armoured vehicle with a likely layout for its daylight and night time vision[5]. For small arms a simpler system would be adequate to meet the user requirements[6]. Figure 9.9 shows the photograph of one such system developed by the Defence R&D Organisation for the Indian Army[5]. REFERENCES 1.

2. 3. 4. 5.

6.

Blackler,F.G., "Practical Guide to Night Viewing System Performance", SPIE, Assessment of Imaging System, Visible and Infrared (SIRA). vol. 274, (1981), pp.248-55. Soule, H.V. "Electro-optical Photography at Low Illumination Levels". (John Wiley and Sons Inc). Report on Creation of Test Facilities for Night Vision in the 300ft Long Hall. (IRDE, Dehradun). Integrated Test Equipment for Night Vision Devices. (Perfect Electronics, Dehradun). Six Photographs and One Sketch. Various Night Vision Devices. (Courtesy, Instruments Research & Development Establishment, Dehradun). Gourley, S & Henish, M. "Sensors for Small Arms". International Defence Review. vol. 5, 1995, pp. 53-57.

161

A

Index

Accommodation 1 Acquisition 16, 23 Active imaging 56 Airy disc 66, 68, 69 Alloy Photocathodes 81 Aperture lens, 118 Aperture stop 62 Atmospheric windows 30, 57 Attenuation coefficient 33

B Back focal-length 61 Background 44 Bipotential lens 118 Blackwell’s Approach 18

C Cathodo-luminescence 93 Charge coupled devices (CCD) 14, 133 Collimator type test equipment 154 Composite Photocathodes 81 Cone Photoreceptors 11 Cone Receptors 6 Cones 9 Contrast 23, 35, 145

E

Electron image intensifier 120 Electron-optics 117 Entrance pupil 62 Environment 29 Exit pupil 62 Experimental lab testing for range-evaluation 150 Eyepieces 59

F Fibre-optic twisters 109 Fibre-optics 109 Field of view 61 Field stop 62 Field test 154 Focal-length 61 Front focal-length 61 Fused fibre-optics faceplates 109

G Generation-0 125 night vision system 155 Generation-1 106, 125 Generation–2 106, 109 Wafer Tube 129 Generation-3 107, 131

D

H

Detection 17, 20, 23 of movement 22 probability 23

Hand-held night vision binocular of the Generation-2 156

164

An Introduction to Night Vision Technology

Human eye 2 Hybrid tubes 131

I Image Intensifier based night vision observation device 156 Identification 17, 20 Image intensifier tubes Production of 138 Performance 134 Signal–to–noise ratio 134 Image intensifiers 14 Function 149 Imaging performance 26 Instrument systems 143 Instrument types 155

J Johnson criteria 19, 23

L Langrange-helmholtz equation 119 Low light level television system 157 Luminescence decay 99 Luminous gain & E.B.I 137 Luminous sensitivity 89

M Magnification 60, 63 Micro-bolometer 57 Modulation transfer function 66, 69, 24, 136 Monochromatic diffraction limited mtf 66, 67, 68 Moonlight 44, 46 Multi-alkali photocathode 140

N Natural backgrounds 50 Night illumination 43

Night time sniper’s rifle telescope 161 Night vision devices 52 Image intensifier function 149 Night vision system 147 Optical factor 148 Night-time turbulence 40 Noise power factor 146 Numerical aperture 63

O Object scene factor 148 Optic flow 1 Optical designs 55 Optical parameters Schematic eye 3 Optical system 59 Orientation 20

P Paraxial approach 64 Passive imaging 56 Perfect image 58 Perfect optical system 58 Phosphor Luminous transitions in 94 Phosphor ageing 104 Phosphors 93 Luminescence efficiency 99 Photocathode 52, 120 Alkali 82 Dark current in 90 Efficiency of 79 Negative affinity 83 Response time 87 Sensitivity 87 Transferred electron 85 Types 80, 83 Photoemission 77

Q Quantum contrast 147

Index

Quantum detection efficiency 25 Quantum detectors 57 Quantum starved 43

R Range equation 144 Ray tracing 65 Recognition 17, 20 Reflectivity at night 48 Retina 2, 5, 7, 24 Rods 9

S Scattering coefficient in rainfall 34 Schematic eye 2 Screen fabrication 103 Signal-to-noise 27 Signal-to-noise ratio 20 Silicon self-scanning array systems 133 Small arms 161 Snell’s law 64

Snow 34 Starlight 46, 48 Stereopsis 1

T Target 44 Thermal imaging 160 Third order aberrations 64 Transmission 30 Trigonometrical ray tracing 65

U Unipotential lens 118

V Vignetting 62 Visibility 35 Vision 1, 8 Vision cues 17 Visual system 6

Z Zero generation 105

165