The book is divided into six parts. The first part (Chapter 2) analyses information about remote sensing satellites; the second part (Chapters 3 and 4) analyses vegetation and shadow-water indices; the third part (Chapters 5-8) is the most consistent and proposes features for land use classification based on structural, multispectral, hybrid, and graph theoretical methods; the fourth part (Chapters 9 and 10) addresses the problem of detecting residential regions exploiting measures of spatial coherence; the fifth part (Chapters 11 and 12) deals with building and road detection in residential regions by means of a complex approach based on a graph theoretical representation of the balloons extracted from the binary image obtained from a clustering algorithm; the sixth part summarizes the overall performance of their system. In my opinion the review questions at the end of each chapter are very useful, particularly if this book will be used for teaching.

In Chapter 1 (Introduction) the authors write that their aims were the proposal of a novel automated end-to-end system to analyze multispectral satellite images and to emphasize how many research problems in remote sensing applications are waiting to be solved by the computer vision community. Well, the book satisfies both these goals.

Chapter 5 (Review on Land Use Classification) is absolutely the most interesting chapter of the book, primarily because it seems a new part, not found in previous works. Particularly interesting is the taxonomy of the feature extraction methods illustrated Fig. 5.1. In fact, to identify trends in feature extraction methods the authors grouped the over 90 influential papers that they had reviewed, all published in refereed journals, according to this taxonomy. Despite the fact that the authors were unable to compare the performances of these methods directly, because the results given in most of the papers reviewed are based on just one (or a very few) images and it is rare that two papers evaluate the same images, they report the performances of each method to give the reader some idea of their relative classification rate. However, this review presents the following two strong limitations, reported by the authors themselves: 1) the review investigates trends in land use classification between years 1967 and 2002 only; 2) they did not attempt to cover the whole literature but focused on feature extraction methods using passive sensors and excluded works on classifiers, neural networks, and fuzzy logic.

This brings me to more critical comments. My reading of this book was less exciting than I expected. First, because, as a specialist in (aerial and) satellite multispectral image understanding (up to the point of founding a university spin-off company for the technological transfer of a software developed for the automated land cover, land use mapping), I already knew all the journal papers of the authors of this book, who are among the leading experts in the field. My second disappointment in reading this book was that I did not receive any help in overcoming what I consider a great difficulty for researchers in remote sensing applications, that is to retrieve results coming from many different scientific communities.

References in this book deserve two special comments. First, all references, with the exception of authors’ ones, are at least ten years old, so that the book can be considered out of date under this point of view. Second, I do not know why Springer accepted the publication of a book without reporting in the References the titles of papers cited, but this is a very strong limitation in my opinion, especially in review-chapters like 5 and 11 in this book.

In conclusion, the book is not suitable for experts in the field (who should already know the techniques here described), nevertheless it represents a good reference book, even a milestone, for teaching multispectral image understanding to students and/or young researchers. The worst aspect, considering the rapid development of remote sensing technology in recent years, remains the delay of at least five years in the publication date, with a very negative impact on References.


[postscriptum: Chapters and their journal versions]

Chapter 3 (Linearized Vegetation Indices) is a reprint of “Linearized vegetation indices based on a formal statistical framework; Ünsalan, C.; Boyer, K.L.; IEEE Trans. on Geoscience and Remote Sensing, 42(7), 2004, 1575 – 1585. Chapter 4 (Linearized Shadow and Water Indices) is also derived from the above journal publication (Section VI), but in this case the content is partially rewritten, due to the use of regions from Indiana and Florida instead of New Mexico and Minnesota, in addition to Maryland and Oregon.

Chapter 6 (Land Use Classification using Structural Features) is the reprint of part of “Classifying land development in high-resolution panchromatic satellite images using straight-line statistics”; Ünsalan, C.; Boyer, K.L.; IEEE Trans. on Geoscience and Remote Sensing, 42(4), 2004, 907 – 919 (indeed, the references to their previous works at the beginning of the Chapter 3 and 6 must be reversed). While the rest of the above journal paper is reprinted in Chapter 9 (Feature Based Grouping to Detect Suburbia): all its six figures and paragraphs 9.1 and 9.2 of the book are derived from Section IV and V.C of the paper, respectively.

Chapter 7 (Land Use Classification via Multispectral Information) is a reprint, starting from section IV of the paper to its end, of “Classifying land development in high-resolution Satellite imagery using hybrid structural-multispectral features”; Ünsalan, C.; Boyer, K.L.; IEEE Trans. on Geoscience and Remote Sensing, 42(12), 2004, 2840 – 2850.

Chapter 8 (Graph Theoretical Measures for Land Development) is a reprint of a subset of “A theoretical and experimental investigation of graph theoretical measures for land development in satellite imagery”; Ünsalan, C.; Boyer, K.L.; IEEE Trans. on PAMI, 27(4), 2005, 575 – 589. While the rest of the above journal paper is reprinted in Chapter 10 (Detecting Residential Regions by Graph-Theoretical Measures): all its tables and paragraphs 10.1 and 10.2 of the book are derived from Section 6.4.1 and 6.4.2 of the paper, respectively. There are not References in Chapter 10.

Chapter 11 (Review on Building and Road Detection) is a brief (5 pages) literature review on building and road detection. The authors cite 43 papers, including the paper “State of the art on automatic road extraction for GIS update: a novel classification”, J.B. Mena, Pattern Recognition Letters 24 (2003) 3037–3058, which reports nearly 250 references related to this topic but (obviously) only till 2003.

Chapter 12 (House and Street Network Detection in Residential Regions) is a reprint, starting from section 2 of the paper to its end, of “A system to detect houses and residential street networks in multispectral satellite images”; Ünsalan, C.; Boyer, K.L.; Computer Vision and Image Understanding, 98(3), 2005, 423–461.

Click here for Top of Page



Multispectral Satellite Image Understanding



Cem Ünsalan and Kim L. Boyer


Series:  Advances in Computer Vision and Pattern Recognition

Springer, 2011


Reviewed by

Primo Zingaretti (Italy)

Recent book reviews:


Multispectral Satellite Image Understanding

By Cem Ünsalan and Kim L. Boyer

(reviewed in this issue)


Automatic Digital Document

Processing and Management: 

Problems, Algorithms, and Techniques

By Stefano Ferilli

(reviewed in this issue)


Automatic Calibration and Reconstruction for Active Vision Systems

By Beiwei Zhang and Y. F. Li

(reviewed in this issue)


Fundamentals of Digital Image Processing: 

A Practical Approach with Examples in Matlab

by Chris Solomon and Toby Breckon

Apr ‘12 [html]    [pdf]


An Introduction to Biometrics

by Anil Jain, Arun Ross,

and Karthik Nandhakumar

Jan ‘12 [html]    [pdf]


Handbook of Geometric Computing

by Eduardo Bayro Corrachano (Ed.)

Jul ‘11 [html]     [pdf]


Essential Image Processing and GIS for

Remote Sensing

by Jian Guo Liu and Philippa Mason

Jul ‘11 [html]     [pdf]


Handbook of Pattern Recognition and

Computer Vision, 4th Edition

by C.H. Chen

Jul ‘11 [html]     [pdf]

Right Arrow: Previous
Right Arrow: Next