Silva

BOOK REVIEW

Remote Sensing and Urban Analysis Edited by: Jean-Paul Donnay, Mike J. Barnsley, and Paul A. Longley (GISDATA 9, Series Editors: Ian Masser and François Salgé) (London, UK:Taylor and Francis), 2001. 268 pages.

The use of remote sensing in urban analysis presents multiple problems when compared with image classification in natural areas. The size of the pixel, the mixture of different land covers, and the irregularity of shapes ensure the difficulty of per-pixel classification methods. This explains the difficulty of defining land cover and use inside the urban boundary and justifies some of the lack of interaction between professionals in urban planning and analysis and those in remote sensing. In light of this gap, Remote Sensing and Urban Analysis is a welcome addition to the scarce literature on urban analysis and remote sensing. The publication of this book was motivated by the recent wave of high-resolution satellites (<5 meters) and the assumption that there is a need for more interaction between professionals involved in the urban realm. The book has its roots in a specialist meeting on remote sensing and urban analysis held at the University of Strasbourg, France and sponsored by the European Science Foundation’s (ESF) GISDATA Program. The intent of the book is to unveil what modern digital remote sensing can offer both urban planners and society in general. The method is to explore the potential of the technology in general, and remote sensing in particular. Remote Sensing and Urban Analysis is an assemblage of superbly written chapters on theory and practice. It is a resource for remote sensing, urbanism, geography, and planning, and includes contributions from many prominent experts on remote sensing, urban analysis, and GIS. It features themes that range from the continuous need to advance image processing, to go beyond the simple mapping of the physical form of urban areas, URISA Journal • Book Review

to conclude the importance of defining new models and spatial analysis tools, and to the improvement of the communication between the different specialists. The authors are effective at catching our attention with the broad idea, exploring the multiple implications, and then detailing their contributions in a coherent manner. The book is a reflective and well-written piece of work that is needed and that highlights the new opportunities available for students and professionals, as a doubt taker for those working on a daily basis and an important landmark in the advent of this third generation of very high spatial resolution of satellite sensors. The book is divided into four parts, and the topics intertwine without losing the flow and the attention of the reader. The authors seem to speak with each other throughout the chapters, sharing problems, acknowledgments, and doubts, as well as making this a comprehensive source. Part 1 begins with an Introduction by Donnay, Barnsley, and Longley that details the major issues to be discussed. The state of the art in urban remote sensing, the demand for higher spatial resolution data, the enhancement of the classification process, and the importance of going from urban morphology/physical dimension to urban/human dimension are some of the subjects presented. Ranchin et al. (Chapter 2) explored the fact that even though a new generation of sensors is available, other methods are available for merging old and new datasets. He explores a method based on the wavelet transform. Brvio et al. (Chapter 3) uses geostatistical techniques of fitting parameters resulting from semivarigrams to the data sets to distinguish different urban en33

vironments. Bianchin (Chapter 4) explores the use of mathematical morphology operators to distinguish urban areas and their form. Part 2 devotes considerable attention to the use of traditional per-pixel classification based in spectral signature and important new techniques such as image segmentation (broadly accepted by the authors as one of best techniques, but still not widely used). It begins with a chapter by Mesev (Chapter 5) that reviews the problems associated with the more traditional standard techniques of statistical classification and a number of practical ways to overcome those problems. Bahr (Chapter 6) explores the importance of segmentation techniques. Barnsley (Chapter 7) takes it further, detailing the techniques of structural pattern recognition. It flows from the readings that the identification of the morphological properties of urban areas is hardly identifiable through spectral signature and per-pixel analysis alone. Urban complexity implies important relations of neighborhood that cannot be ignored. As soon as we understand the importance of urban morphology and its different facets, the authors open another chapter that goes beyond. The importance of the urban/human dimension, the possibility of using remote sensing to infer the character of the place, and the possibility of challenging theories and methods are raised. Part IV (“Defining Urban Populations Over Space and Time”) begins with Longly and Messev (Chapter 9) addressing the importance of urban morphologies described in the previous chapter to derive indicators of urban shape and form. With that comes the questioning of data-led theories leading to theory-led data analysis. Issues such as fractal dimension/ allometry should be part of our urban agendas. Batty and Howes (Chapter 10) explain how time is included in remote sensing and urban analysis; they seek to understand and predict temporal patterns, using a method to classify development by age with remote sensing and ancillary data. Using time discontinuities regarding differences in densities, building materials, and transport systems, three important time frames are proposed: the 1930s, 1960s, and 1990s (p. 187). Donnay and Unwin (Chapter 11) direct our attention to the possibility of disclosing different population densities, also with the help of ancillary data. Finally, Baudout (Chapter 12) puts geography into perspective: do not forget that different continents have different realities, different objectives, and certainly different methods. This sincere and clear disclosure of information throughout the book telling us of what we need to be aware or the unfolding opportunities that we will find along the way makes this book interesting.

34

Chapter 13, “A Research Agenda in Remote Sensing and Urban Analysis,” is a highlight of the book. As the editors point out, it is important to challenge remote sensing analysis, by going beyond per-pixel analysis into questions of configuration, syntax, structure, and function. Contrary to what is stated in the Epilogue, the book did not come out too early (p. 245), but just in time! It prepares the readers for what we should expect (potentialities, problems, advantages, questions still unsolved) in different opinions, methods, and how the different fields need to be related. As the authors conclude, there is still much to do on some real problems. The possibilities of this new wave of satellites also presages new problems, and there is a danger that remote sensing specialists will expend more effort addressing the opportunities and problems posed by each new technological advance, rather than seeking operational solutions to the use of existing systems, mainly if we want to go further than the simple urban form and try to extract other functional properties that are to the interest of urban planners and cannot be observed directly (p. xxiii). Much needs to be done in terms of the outdateness of some ancillary data to monitor fast changing urban development trends, or the balance between coarse resolution and confidentiality, and the incompleteness of some of the data sets (p. 252). These are presented and explored as a tool to help us realize what needs to be done. The authors see solutions through collaboration, with this book as proof. As stated in the Conclusion “the outcome of research will determine whether urban remote sensing is to fulfill only a technical role in corroborating and updating other data sources, or whether it might fulfill a more central role in terms of quality of life studies, and data-rich modeling of form and function”(p. 255). Reading this book reveals the encouraging optimism of the authors. Overall, the aim of this book is well accomplished. It is indeed a balanced exposition of detailed and broad views, and the bibliography helps the reader who wants to go further. Reviewed by: Elisabete A. Silva Department of Landscape Architecture and Regional Planning University of Massachusetts Amherst, Massachusetts [email protected]

URISA Journal • Vol. 13, No. 2 • Spring 2001