Change Detection and Image Time Series Analysis 1 Unsupervised Methods 1St Edition Abdourrahmane M Atto Full Chapter PDF

Change Detection and Image Time
Series Analysis 1: Unsupervised

Methods 1st Edition Abdourrahmane M.
Atto
Visit to download the full and correct content document:
https://ebookmass.com/product/change-detection-and-image-time-series-analysis-1-u
nsupervised-methods-1st-edition-abdourrahmane-m-atto/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Methods and techniques for fire detection : signal,

image and video processing perspectives 1st Edition
Çetin
https://ebookmass.com/product/methods-and-techniques-for-fire-
detection-signal-image-and-video-processing-perspectives-1st-
edition-cetin/
Interrupted Time Series Analysis David Mcdowall
https://ebookmass.com/product/interrupted-time-series-analysis-
david-mcdowall/
Applied economic forecasting using time series methods

Ghysels
https://ebookmass.com/product/applied-economic-forecasting-using-
time-series-methods-ghysels/
Time Series Analysis in Seismology: Practical

Applications 1st Edition Alejandro Ramírez-Rojas
https://ebookmass.com/product/time-series-analysis-in-seismology-
practical-applications-1st-edition-alejandro-ramirez-rojas/
Time Villains Series, Book 1 Victor Piñeiro
https://ebookmass.com/product/time-villains-series-book-1-victor-
pineiro/
Demanding Energy: Space, Time and Change 1st Edition

Allison Hui
https://ebookmass.com/product/demanding-energy-space-time-and-
change-1st-edition-allison-hui/
Machine vision inspection systems. Vol.1: image

processing, concepts Malarvel M (Ed.)
https://ebookmass.com/product/machine-vision-inspection-systems-
vol-1-image-processing-concepts-malarvel-m-ed/
Deep Learning for Medical Image Analysis 1st Edition S.

Kevin Zhou
https://ebookmass.com/product/deep-learning-for-medical-image-
analysis-1st-edition-s-kevin-zhou/
Applied Time Series Analysis. A Practical Guide to

Modeling and Forecasting Terence C. Mills
https://ebookmass.com/product/applied-time-series-analysis-a-
practical-guide-to-modeling-and-forecasting-terence-c-mills/
Change Detection and Image Time Series Analysis 1
SCIENCES
Image, Field Director – Laure Blanc-Feraud
Remote Sensing Imagery, Subject Heads –
Emmanuel Trouvé and Avik Bhattacharya
Change Detection and Image

Time Series Analysis 1
Unsupervised Methods
Coordinated by
Abdourrahmane M. Atto
Francesca Bovolo
Lorenzo Bruzzone
First published 2021 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced,
stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentioned address:
ISTE Ltd John Wiley & Sons, Inc.

27-37 St George’s Road 111 River Street
London SW19 4EU Hoboken, NJ 07030
UK USA
www.iste.co.uk www.wiley.com
© ISTE Ltd 2021

The rights of Abdourrahmane M. Atto, Francesca Bovolo and Lorenzo Bruzzone to be identified as the
authors of this work have been asserted by them in accordance with the Copyright, Designs and Patents
Act 1988.
Library of Congress Control Number: 2021941648
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78945-056-9
ERC code:
PE1 Mathematics
PE1_18 Scientific computing and data processing
PE10 Earth System Science
PE10_3 Climatology and climate change
PE10_4 Terrestrial ecology, land cover change
PE10_14 Earth observations from space/remote sensing
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Abdourrahmane M. ATTO, Francesca B OVOLO and Lorenzo B RUZZONE
List of Notations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Chapter 1. Unsupervised Change Detection in Multitemporal

Remote Sensing Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Sicong L IU, Francesca B OVOLO, Lorenzo B RUZZONE, Qian D U
and Xiaohua T ONG
1.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Unsupervised change detection in multispectral images . . . . . . . . . 3
1.2.1. Related concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.2. Open issues and challenges . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.3. Spectral–spatial unsupervised CD techniques . . . . . . . . . . . . . 7
1.3. Unsupervised multiclass change detection approaches based on
modeling spectral–spatial information . . . . . . . . . . . . . . . . . . . . . . 9
1.3.1. Sequential spectral change vector analysis (S2 CVA) . . . . . . . . . 9
1.3.2. Multiscale morphological compressed change vector analysis . . . 11
1.3.3. Superpixel-level compressed change vector analysis . . . . . . . . . 15
1.4. Dataset description and experimental setup . . . . . . . . . . . . . . . . 18
1.4.1. Dataset description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.4.2. Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.5. Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.5.1. Results on the Xuzhou dataset . . . . . . . . . . . . . . . . . . . . . 24
1.5.2. Results on the Indonesia tsunami dataset . . . . . . . . . . . . . . . 24
vi Change Detection and Image Time Series Analysis 1
1.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.7. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.8. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Chapter 2. Change Detection in Time Series of Polarimetric

SAR Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Knut C ONRADSEN, Henning S KRIVER, Morton J. C ANTY
and Allan A. N IELSEN
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.1.1. The problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.1.2. Important concepts illustrated by means of the gamma
distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.2. Test theory and matrix ordering . . . . . . . . . . . . . . . . . . . . . . . 45
2.2.1. Test for equality of two complex Wishart distributions . . . . . . . . 45
2.2.2. Test for equality of k-complex Wishart distributions . . . . . . . . . 47
2.2.3. The block diagonal case . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.2.4. The Loewner order . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.3. The basic change detection algorithm . . . . . . . . . . . . . . . . . . . 53
2.4. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.4.1. Visualizing changes . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.4.2. Fieldwise change detection . . . . . . . . . . . . . . . . . . . . . . . 59
2.4.3. Directional changes using the Loewner ordering . . . . . . . . . . . 62
2.4.4. Software availability . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.5. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Chapter 3. An Overview of Covariance-based Change Detection

Methodologies in Multivariate SAR Image Time Series . . . . . . . . 73
Ammar M IAN, Guillaume G INOLHAC, Jean-Philippe OVARLEZ,
Arnaud B RELOY and Frédéric PASCAL
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.2. Dataset description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.3. Statistical modeling of SAR images . . . . . . . . . . . . . . . . . . . . 77
3.3.1. The data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.2. Gaussian model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.3.3. Non-Gaussian modeling . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.4. Dissimilarity measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.1. Problem formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.4.2. Hypothesis testing statistics . . . . . . . . . . . . . . . . . . . . . . . 85
3.4.3. Information-theoretic measures . . . . . . . . . . . . . . . . . . . . . 87
3.4.4. Riemannian geometry distances . . . . . . . . . . . . . . . . . . . . 89
3.4.5. Optimal transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Contents vii
3.4.6. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.4.7. Results of change detectors on the UAVSAR dataset . . . . . . . . . 91
3.5. Change detection based on structured covariances . . . . . . . . . . . . 94
3.5.1. Low-rank Gaussian change detector . . . . . . . . . . . . . . . . . . 96
3.5.2. Low-rank compound Gaussian change detector . . . . . . . . . . . . 97
3.5.3. Results of low-rank change detectors on the UAVSAR dataset . . . 100
3.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
3.7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Chapter 4. Unsupervised Functional Information Clustering

in Extreme Environments from Filter Banks and
Relative Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Abdourrahmane M. ATTO, Fatima K ARBOU, Sophie G IFFARD -ROISIN
and Lionel B OMBRUN
4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.2. Parametric modeling of convnet features . . . . . . . . . . . . . . . . . . 110
4.3. Anomaly detection in image time series . . . . . . . . . . . . . . . . . . 113
4.4. Functional image time series clustering . . . . . . . . . . . . . . . . . . 119
4.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.6. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Chapter 5. Thresholds and Distances to Better Detect Wet Snow

over Mountains with Sentinel-1 Image Time Series . . . . . . . . . . . 127
Fatima K ARBOU, Guillaume JAMES, Philippe D URAND
and Abdourrahmane M. ATTO
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.2. Test area and data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.3. Wet snow detection using Sentinel-1 . . . . . . . . . . . . . . . . . . . . 129
5.4. Metrics to detect wet snow . . . . . . . . . . . . . . . . . . . . . . . . . 133
5.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.8. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Chapter 6. Fractional Field Image Time Series Modeling

and Application to Cyclone Tracking . . . . . . . . . . . . . . . . . . . . 145
Abdourrahmane M. ATTO, Aluísio P INHEIRO, Guillaume G INOLHAC
and Pedro M ORETTIN
6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.2. Random field model of a cyclone texture . . . . . . . . . . . . . . . . . 148
viii Change Detection and Image Time Series Analysis 1
6.2.1. Cyclone texture feature . . . . . . . . . . . . . . . . . . . . . . . . . 149

6.2.2. Wavelet-based power spectral densities and cyclone fields . . . . . . 150
6.2.3. Fractional spectral power decay model . . . . . . . . . . . . . . . . . 153
6.3. Cyclone field eye detection and tracking . . . . . . . . . . . . . . . . . . 157
6.3.1. Cyclone eye detection . . . . . . . . . . . . . . . . . . . . . . . . . . 157
6.3.2. Dynamic fractal field eye tracking . . . . . . . . . . . . . . . . . . . 158
6.4. Cyclone field intensity evolution prediction . . . . . . . . . . . . . . . . 159
6.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
6.7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Chapter 7. Graph of Characteristic Points for Texture Tracking:

Application to Change Detection and Glacier Flow Measurement
from SAR Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Minh-Tan P HAM and Grégoire M ERCIER
7.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
7.2. Texture representation and characterization using local extrema . . . . 169
7.2.1. Motivation and approach . . . . . . . . . . . . . . . . . . . . . . . . 169
7.2.2. Local extrema keypoints within SAR images . . . . . . . . . . . . . 172
7.3. Unsupervised change detection . . . . . . . . . . . . . . . . . . . . . . . 175
7.3.1. Proposed framework . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
7.3.2. Weighted graph construction from keypoints . . . . . . . . . . . . . 176
7.3.3. Change measure (CM) generation . . . . . . . . . . . . . . . . . . . 178
7.4. Experimental study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.4.1. Data description and evaluation criteria . . . . . . . . . . . . . . . . 179
7.4.2. Change detection results . . . . . . . . . . . . . . . . . . . . . . . . . 181
7.4.3. Sensitivity to parameters . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.4.4. Comparison with the NLM model . . . . . . . . . . . . . . . . . . . 188
7.4.5. Analysis of the algorithm complexity . . . . . . . . . . . . . . . . . 191
7.5. Application to glacier flow measurement . . . . . . . . . . . . . . . . . 192
7.5.1. Proposed method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
7.5.2. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
7.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
7.7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Chapter 8. Multitemporal Analysis of Sentinel-1/2 Images

for Land Use Monitoring at Regional Scale . . . . . . . . . . . . . . . . 201
Andrea G ARZELLI and Claudia Z OPPETTI
8.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
8.2. Proposed method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
8.2.1. Test site and data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Contents ix
8.3. SAR processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

8.4. Optical processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
8.5. Combination layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
8.6. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
8.7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
8.8. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Chapter 9. Statistical Difference Models for Change Detection

in Multispectral Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Massimo Z ANETTI, Francesca B OVOLO and Lorenzo B RUZZONE
9.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
9.2. Overview of the change detection problem . . . . . . . . . . . . . . . . 225
9.2.1. Change detection methods for multispectral images . . . . . . . . . 227
9.2.2. Challenges addressed in this chapter . . . . . . . . . . . . . . . . . . 230
9.3. The Rayleigh–Rice mixture model for the magnitude of the
difference image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
9.3.1. Magnitude image statistical mixture model . . . . . . . . . . . . . . 231
9.3.2. Bayesian decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
9.3.3. Numerical approach to parameter estimation . . . . . . . . . . . . . 234
9.4. A compound multiclass statistical model of the difference image . . . . 239
9.4.1. Difference image statistical mixture model . . . . . . . . . . . . . . 240
9.4.2. Magnitude image statistical mixture model . . . . . . . . . . . . . . 245
9.4.3. Bayesian decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
9.4.4. Numerical approach to parameter estimation . . . . . . . . . . . . . 249
9.5. Experimental results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
9.5.1. Dataset description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
9.5.2. Experimental setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
9.5.3. Test 1: Two-class Rayleigh–Rice mixture model . . . . . . . . . . . 256
9.5.4. Test 2: Multiclass Rician mixture model . . . . . . . . . . . . . . . . 260
9.6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
9.7. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Summary of Volume 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

Preface
Abdourrahmane M. ATTO1 , Francesca B OVOLO2

and Lorenzo B RUZZONE3
1
University Savoie Mont Blanc, Annecy, France
2
Fondazione Bruno Kessler, Trento, Italy
3
University of Trento, Italy
This book is part of the ISTE-Wiley “SCIENCES” Encyclopedia and belongs to

the Image field of the Engineering and Systems department. The Image field covers
the entire processing chain from acquisition to interpretation by analyzing the data
provided by various imaging systems. This field is split into seven subjects, including
Remote Sensing Imagery (RSI). The heads of this subject are Emmanuel Trouvé and
Avik Bhattacharya. In this subject, we propose a series of books that portray diverse
and comprehensive topics in advanced remote-sensing images and their application
for Earth Observation (EO). There has been an increasing demand for monitoring and
predicting our planet’s evolution on a local, regional and global scale. Hence, over
the past few decades, airborne, space-borne and ground-based platforms with active
and passive sensors acquire images that measure several features at various spatial and
temporal resolutions.
RSI has become a broad multidisciplinary domain attracting scientists across the
diverse fields of science and engineering. The aim of the books proposed in this RSI
series is to present the state-of-the-art and available scientific knowledge about the
primary sources of images acquired by optical and radar sensors. The books cover
the processing methods developed by the signal and image processing community to
extract useful information for end-users for an extensive range of EO applications in
natural resources.
In this project, each RSI book focuses on general topics such as change
detection, surface displacement measurement, target detection, model inversion and
Change Detection and Image Time Series Analysis 1,

coordinated by Abdourrahmane M. ATTO, Francesca B OVOLO and Lorenzo B RUZZONE .
© ISTE Ltd 2021.
xii Change Detection and Image Time Series Analysis 1
data assimilation. This first book of the RSI series is dedicated to Change Detection
and Image Time Series Analysis. It presents methods developed to detect changes
and analyze their temporal evolutions using optical and/or synthetic aperture radar
(SAR) images in diverse settings (e.g. image pairs, image time series). According
to the numerous works and applications in this domain, this book is divided into
two volumes, dedicated to unsupervised and supervised approaches, respectively.
Unsupervised methods require little to no expert-based information to resolve a
problem, whereas the contrary holds true, especially for methods that are supervised
in the sense of providing a wide amount of labeled training data to the method, before
testing this method.
Volume 1: Unsupervised methods
A significant part of this book is dedicated to a wide range of unsupervised

methods. The first chapter provides an insight into the motivations of this behavior and
introduces two unsupervised approaches to multiple-change detection in bitemporal
multispectral images. Chapters 2 and 3 introduce the concept of change detection
in time series and postulate it in the context of statistical analysis of covariance
matrices. The former chapter focuses on a directional analysis for multiple-change
detection and exercises on a time series of SAR polarimetric data. The latter focuses
on local analysis for binary change detection and proposes several covariance matrix
estimators and their corresponding information-theoretic measures for multivariate
SAR data. The last four chapters focus more on applications. Chapter 4 addresses
functional representations (wavelets and convolutional neural network filters) for
feature extraction in an unsupervised approach. It proposes anomaly detection
and functional evolution clustering from this framework by using relative entropy
information extracted from SAR data decomposition. Chapter 5 deals with the
selection of metrics that are sensitive to snow state variation in the context of
the cryosphere, with a focus on mountain areas. Metrics such as cross-correlation
ratios and Hausdorff distance are analyzed with respect to optimal reference images
to identify optimal thresholding strategies for the detection of wet snow by using
Sentinel-1 image time series. Chapter 6 presents time series analysis in the context of
spatio-temporal forecasting and monitoring fast-moving meteorological events such
as cyclones. The application benefits from the fusion of remote sensing data under the
fractional dynamic field assumption on the cyclone behavior. Chapter 7 proposes an
analysis based on characteristic points for texture modeling with graph theory. Such
an approach overcomes issues arising from large-size dense neighborhoods that affect
spatial context-based approaches. The application proposed in this chapter concerns
glacier flow measurement in bitemporal images. Chapter 8 focuses on detecting
new land-cover types by classification-based change detection or feature/pixel-based
change detection. Monitoring the construction of new buildings in urban and suburban
scenarios at a large regional scale by means of Sentinel-1 and -2 images is considered
as an application. Chapter 9 focuses on the statistical modeling of classes in the
Preface xiii
difference image and derives from scratch a multiclass model for it in the context
of change vector analysis.
Volume 2: Supervised methods
The second volume of this book is dedicated to supervised methods. Chapter 1 of

this volume addresses the fusion of multisensor, multiresolution and multitemporal
data. This chapter reviews recent advances in the literature and proposes two
supervised Markov random field-based solutions: one relies on a quadtree and the
second one is specifically designed to deal with multimission, multifrequency and
multiresolution time series. Chapter 2 provides an overview of pixel-based methods
for time series classification from the earliest shallow-learning methods to the most
recent deep learning-based approaches. This chapter also includes best practices for
reference data preparation and management, which are crucial tasks in supervised
methods. Chapter 3 focuses on very high spatial resolution data time series and the use
of semantic information for modeling spatio-temporal evolution patterns. Chapter 4
focuses on the challenges of dense time series analysis, including pre-processing
aspects and a taxonomy of existing methodologies. Finally, since the evaluation of
a learning system can be subject to multiple considerations, Chapters 5 and 6 propose
extensive evaluations of the methodologies used to produce earthquake-induced
change maps, with an emphasis on their strengths and shortcomings (Chapter 5)
and the deep learning systems in the context of multiclass multilabel change-of-state
classification on glacier observations (Chapter 6).
This book covers both methodological and application topics. From the
methodological viewpoint, contributions are provided with respect to feature
extraction and a large number of evaluation metrics for change detection, classification
and forecasting issues. Analysis has been performed in both bitemporal images and
time series, illustrating both unsupervised and supervised methods and considering
both binary- and multiclass outputs. Several applications are mentioned in the
chapters, including agriculture, urban areas and cryosphere analysis, among others.
This book provides a deep insight into the evolution of change detection and time
series analysis in the state-of-the-art, as well as an overview of the most recent
developments.
July 2021
List of Notations
I “ pIk pp, qqqk,p,q Image Time Series: time index k and pixel position
pp, qq
I “ pÍkc pp, qqqk,p,q,c
¯
Vector Image Time Series: band/spectral index c
pu,vq
I “ Ik pp, q q Matrix Image Time Series: (polarimetric indices pu, v q)
k,p,q,u,v
N, Z, R, C Sets of Natural Numbers, Integers, Real and Complex
Numbers
μ, μ Means of Random Variables and Random Vectors
C, Σ Physical and Statistical Variance–Covariance Matrices
pdf Probability Density Function

© ISTE Ltd 2021.
1
Unsupervised Change Detection
in Multitemporal Remote
Sensing Images
Sicong L IU1 , Francesca B OVOLO2 , Lorenzo B RUZZONE3 ,

Qian D U4 and Xiaohua T ONG1
1
Tongji University, Shanghai, China
2
Fondazione Bruno Kessler, Trento, Italy
3
University of Trento, Italy
4
Mississippi State University, Starkville, USA
1.1. Introduction
Remote sensing satellites have a great potential to recurrently monitor the

dynamic changes of the Earth’s surface in a wide geographical area, and contribute
substantially to our current understanding of the land-cover and land-use changes
(Bruzzone and Bovolo 2013; Song et al. 2018; Liu et al. 2019c). Scientifically
understanding land changes is also essential for analyzing environmental evolution
and anthropic phenomena, especially when studying the global change and its impact
on human society. Thanks to the satellite revisit property, both long-term (e.g.
yearly) and short-term (e.g. daily) satellite observations produce a huge amount of
multitemporal images in the data archive (Liu et al. 2015). Based on the analysis of
multitemporal data, land-cover changes can be automatically discovered and detected,
where knowledge of changes can also be acquired. This becomes an important
and complementary way of optimizing the traditional in situ investigation, which

© ISTE Ltd 2021.
Change Detection and Image Time Series Analysis 1: Unsupervised Methods,
First Edition. Abdourrahmane M. Atto; Francesca Bovolo and Lorenzo Bruzzone.
© ISTE Ltd 2021. Published by ISTE Ltd and John Wiley & Sons, Inc.
2 Change Detection and Image Time Series Analysis 1
is often very costly and labor-intensive. In particular, in some cases such as in

a natural disaster scenario, it is very difficult or even impossible to conduct field
investigations. Change detection (CD) is the process of identifying change and
no-change regions on the Earth’s surface by analyzing images acquired from the same
geographical area at different times (Singh 1989; Coppin et al. 2004; Lu et al. 2004;
Liu et al. 2019c). Therefore, automatic and robust techniques need to be designed to
effectively discover, describe and detect changes that occur in multitemporal remote
sensing images. In the past decades, CD has become an increasingly active research
field and has been widely applied in various remote sensing applications, such as
deforestation, disaster evaluation, urban expansion evolution and environment and
ecosystem monitoring (Bovolo and Bruzzone 2007a; Bouziani et al. 2010; Du et al.
2012; Khan et al. 2017; Liu et al. 2020b).
Basically, CD techniques are developed based on specific remote sensing satellite

sensors. In the literature, many excellent articles focused on the discussion of CD
problems in different types of satellite sensors: for example, CD in multispectral
images (Lu et al. 2004; Ban and Yousif 2016), SAR images (Ban and Yousif 2016)
and hyperspectral images (Liu et al. 2019c), as well as in Lidar data (Okyay et al.
2019). Among different sensors mounted on the EO satellites, multispectral scanners
can acquire images with both high spatial resolution and wide spatial coverage. In
the past decades, due to data availability, multispectral remote sensing images such as
Landsat and Sentinel serials contributed the main data source for the Earth’s surface
monitoring and CD applications (Du et al. 2013; Liu et al. 2020a). However, with
the increasing high quality and spatial resolution in new multispectral sensor images,
especially for very high resolution (VHR) images, it is necessary to design advanced
CD techniques that can deal with more complex change patterns presented in a more
complex CD scenario.
In recent years, many CD methods have been developed for multispectral images,
most of which focus on improving the automation, accuracy and applicability of CD
(Leichtle et al. 2017; Liu et al. 2017a, 2019b; Wang et al. 2018; Saha et al. 2019;
Wei et al. 2019). In general, according to the automation degree, they can be grouped
into three main categories: supervised, semi-supervised and unsupervised methods.
Usually, supervised CD approaches have better performance with higher accuracy by
taking advantage of certain robust supervised classifiers (Wang et al. 2018). However,
their implementation relies on the availability of ground reference data, which is
often difficult to collect in most practical cases. Semi-supervised CD approaches start
from limited training samples or partial prior knowledge learned from the single-time
image, where the active learning or transfer learning algorithm can usually be applied
to increase the sample representation (Liu et al. 2017b, 2019a; Zhang et al. 2018; Tong
et al. 2020). In contrast, unsupervised approaches have higher automation without
relying on the availability of ground reference data or prior knowledge (Liu et al.
2017a, 2019b; Saha et al. 2019). Therefore, the analysis in the unsupervised CD
case is mainly data-driven and is actually more challenging than the other two tasks.
Unsupervised Change Detection in Multitemporal Remote Sensing Images 3
However, from a practical application point of view, it is definitely more attractive due
to its simplicity and high automation.
In this chapter, we focus on the unsupervised CD problem in multitemporal

multispectral images. In particular, we investigate and analyze the spectral–spatial
change representation for addressing the important multiclass CD problem. To
this end, two approaches are developed, including a multi-scale morphological
compressed change vector analysis and a superpixel-level multiclass CD. By taking
advantage of the spectral and spatial joint analysis of change information, both
approaches show higher performance than the compared state-of-the-art methods.
Experimental results obtained from two real multispectral datasets confirmed the
effectiveness of the proposed approaches in terms of higher accuracy and efficiency
of CD.
The rest of this chapter is organized as follows. Section 1.2 points out the
key concepts and challenges in unsupervised CD, and especially reviews the
current development of spectral–spatial unsupervised CD techniques. Section 1.3
describes the two proposed unsupervised multiclass CD approaches in detail. Dataset
description and experimental setup are provided in section 1.4. Experimental results
and discussions are present in section 1.5. Finally, section 1.6 draws the conclusion of
this chapter.
1.2. Unsupervised change detection in multispectral images
1.2.1. Related concepts
Depending on the purpose of unsupervised CD tasks, two main categories of

methods are defined: binary change detection and multiclass change detection. The
former aims to separate only the change and no-change classes, whereas the latter
detects changes and distinguishes different classes within the changed pixels. In this
chapter, we consider the latter, which is more attractive but challenging in practical
CD applications. Note that in the unsupervised CD case, no ground truth or prior
knowledge is available, thus the data-driven CD process is more preferable than
the model-driven process. Therefore, the multiclass discrimination represents the
inter-change difference associated with specific land-cover class transitions, whereas
the detailed “from–to” information is absent, making it essentially different from the
supervised case.
In general, the unsupervised CD process includes the following main steps:

(1) multitemporal data pre-processing; (2) feature generation and selection; (3) change
index construction; (4) CD algorithm design; (5) performance evaluation. The main
components of an unsupervised CD are shown in Figure 1.1. Each step is briefly
described and discussed as follows.
4
Multitemporal Data Feature Generation Change Index Change Detection Performance

Pre-processing and Selection Construction Algorithm Design Evaluation
• Calibration • Spectral Features • Differencing • Change Number • Accuracy Index

• Radiometric & • Spectral Indices • Ratioing Estimation • Error Matrix
Atmospheric Corrections • Spatial Features • Distance/Similarity • Thresholding • ROC
• Image Enhancement • Object Features Measurement • Clustering • AUC
• Co-registration • … • Transformation • Deep Networks • Time Cost
• … • … • … • ….
Figure 1.1. The main technical components of an unsupervised CD process
Multitemporal data pre-processing: in this step, different operations such as

calibration, band stripe repair (if any), radiometric and atmospheric corrections, image
enhancement and image-to-image co-registration are usually conducted in order to
generate high-quality pre-processed multitemporal images for CD in the next steps. In
particular, a high precision of co-registration is the core operation for a successful CD,
which may significantly affect the CD performance due to the presence of remaining
residual errors.
Feature generation and selection: features extracted from original multitemporal

images are the critical carrier for representing different characteristics of objects in
the single-time image and their variations in the temporal domain. Features such
as original spectral bands, spectral indices (e.g. Normalized Difference Vegetation
Index – NDVI, Modified Normalized Difference Water Index – MNDWI, Index-based
Built-up Index – IBI) and textures (e.g. mean, contrast, homogeneity) derived from
original bands can be considered in CD. In addition, spatial features generated from
multispectral bands such as wavelet transformation (Celik and Ma 2011), Gabor
filtering (Li et al. 2015), morphological filtering (Falco et al. 2013), etc., provide
important multi-scale geometric information about image objects to improve the
change representation. Recently, deep learning-based CD approaches have shown
great potential in extracting more high-level deep features, which represents a popular
direction in CD research (Mou et al. 2019; Saha et al. 2019).
Change index construction: the change index represents the temporal variations
extracted from multitemporal image features. It can be constructed based on different
operators and algorithms, such as univariate image differencing (Bruzzone and Prieto
2000a), change vector analysis (CVA) (Bovolo and Bruzzone 2007b), ratioing (Bazi
et al. 2005), distance or similarity measures (Du et al. 2012), etc. Transformation
approaches such as iterative reweighted multivariate alteration detection (IR-MAD)
(Nielsen 2007), principal components analysis (PCA) and its kernel version (Nielsen
and Canty 2008; Celik 2009), independent component analysis (ICA) (Liu et al.
2012), are also designed to transform the change information from the original
data space into a projected feature space. However, a careful selection of specific
components representing user-interested changes is required. This is often very
difficult in an unsupervised CD case without prior knowledge about the considered
study area and dataset, which may limit the automation degree of the CD application.
For a summary of the related methods for constructing different types of change index,
readers can refer to the paper by Bovolo and Bruzzone (2015).
Change detection algorithm design: unlike the supervised and semi-supervised

CD methods that rely on the available reference samples, unsupervised CD algorithms
focus more on the automation and accuracy. Thus, basically, most of the unsupervised
CD approaches are data-driven by analyzing the multitemporal data itself. Within this
context, for binary CD, if we consider a given change index generated in the previous
step, for example, the magnitude of differencing image, automatic thresholding such
as empirical segmentation (Bruzzone and Prieto 2000b), Kittler–Illingworth (KI)

(Bazi et al. 2005), Otsu (1979) and Bayesian-based expectation–maximization (EM)
(Bruzzone and Prieto 2000a) are all simple but effective algorithms proposed in the
literature. However, the successful use of such methods depends on the assumption of
a certain data distribution such as Gaussian or Rayleigh–Rice mixture (Bovolo et al.
2012; Zanetti et al. 2015), where a wrong estimation may lead to many detection
errors. On the contrary, clustering algorithms such as k-means, fuzzy c-means and
Gustafson–Kessel clustering (GKC) have been used to address the binary CD problem
(Celik 2009; Ghosh et al. 2011), which are distribution-free but require a specific
setting to avoid unstable performance, such as the accuracy decrease due to random
initialization.
For a multiclass CD case, the unsupervised task becomes more complex since
several sub-problems should be solved simultaneously, including the binary change
and no-change separation, the number of multiclass change estimation and the
multiclass change discrimination (Liu et al. 2019c). In particular, among many
solutions, we recall the classical multiple CD technique – change vector analysis
(CVA) (Malila 1980). It was designed to analyze possible multiple changes in pairs
of bitemporal image bands. A theoretical definition was given to the original CVA
approach in the polar domain to provide a more clear mathematical explanation to
CVA (Bovolo and Bruzzone 2007b). However, it still has a limitation, i.e. only a part
of all possible changes can be detected since only two selected bands are considered
in each implementation. If more spectral channels are considered, it becomes
very difficult to simultaneously model and visualize multidimensional changes. To
break this constraint, a compressed change vector analysis (C2 VA) approach was
proposed, which successfully extended the original CVA to a two-dimensional (2D)
representation of the multi-band problem (Bovolo et al. 2012). Other works in the
literature developed different variations of CVA. For example, a modified CVA
was developed to determine the magnitude threshold and direction by combining
single-date image classification results (Chen et al. 2003). An improved thresholding
approach on change magnitude was designed to optimize the binary separation on
each specific change class (Bovolo and Bruzzone 2011). A hierarchical version of
C2 VA with an adaptive and sequential projection of spectral change vectors (SCVs)
at each level of the hierarchy was proposed to detect multiple changes in bitemporal
hyperspectral images (Liu et al. 2015). In this chapter, we also explore the potential
capability of C2 VA and extend it from the spectral–spatial point of view.
Performance evaluation: similar to the supervised CD methods, unsupervised

binary and multiclass CD approaches can usually be assessed according to the
detection accuracy or error index, such as overall accuracy (OA), Kappa coefficient
(Kappa), omission errors (OE) and commission errors (CE), receiver operating
characteristic (ROC) curve and area under the curve (AUC) value. In this case, the
accuracy evaluation usually relies on the manually interpreted change reference map.
Note that such a reference map is only used for accuracy evaluation, which is not
considered as training data as in the supervised case. In addition, the computational
time cost is also another important indicator that reflects the automation and efficiency
of unsupervised methods.
1.2.2. Open issues and challenges
The current development of unsupervised CD techniques for multispectral remote

sensing images has had great success in many practical applications. However, there
are still open issues and challenges that deserve to be further analyzed, which include
but are not limited to the following:
1) a high-precision multitemporal pre-processing procedure, for example,
co-registration techniques;
2) multitemporal data quality improvement due to bad imaging conditions, such
as system noise, cloud contamination and seasonal spectral variations;
3) advanced techniques for correctly estimating the real number of multiclass
changes in image scenarios;
4) spectral–spatial modeling of change targets to enhance the original pixel-wise
spectral representation;
5) robust and efficient CD approach in an unsupervised fashion, especially for a
large complex CD scene;
6) change feature representation by taking advantage of both machine learning and
deep learning techniques.
1.2.3. Spectral–spatial unsupervised CD techniques
Despite the success of aforementioned CD methods, especially the CVA-based

methods, they mainly focus on the spectral changes in each individual pixel or a local
neighborhood (Bovolo 2009; Bovolo et al. 2012; Liu et al. 2015). The geometrical
characteristics of change targets are not fully modeled and preserved. This may
increase the ambiguity due to abnormal spectral variations in isolated pixels and
errors (e.g. co-registration errors), leading to the presence of omission and commission
errors, especially when dealing with VHR images. In this case, traditional pixel-based
CD methods may lose their effectiveness since they were developed under the
assumption that pixels are spatially independent. However, for multispectral images
in complex urban scenarios, challenging issues may arise due to the limited spectral
representation; thus, the same class of objects may have different spectra, or different
objects may have the same or very similar spectra. This may significantly increase the
detection difficulty, especially when considering the multiclass CD task.
To address the above problems in pixel-based CD (PBCD) techniques,

spectral–spatial joint analysis and object-based CD (OBCD) methods are mainstream
techniques proposed in the literature. For the former, morphological filters (i.e.
self-dual reconstruction filters and alternating sequential filters) were combined with
CVA for binary CD in VHR images (Mura et al. 2008). However, a sliding window
(i.e. structuring element (SE)) for filtering should be fixed at a given level; thus, it is
not robust for multilevel implementation. Morphological attribute profiles (APs) were
applied to extract structure-related geometrical features within the scene from each
date of panchromatic images (Falco et al. 2013). It includes a multilevel extraction
of connected regions in the scene at different scales. Building change information
based on the difference in the multitemporal morphological building index (MBI)
at the feature and decision level was considered for detecting building changes
in VHR images (Huang et al. 2014). A spectral–spatial band expansion approach
was developed to enhance the change representation in multispectral images with
limited bands, where additional bands were generated from both spectral and spatial
viewpoints (Liu et al. 2019b).
For OBCD methods, four main categories exist: image-object, class-object,

multitemporal-object and hybrid CD (Chen et al. 2012; Hussain et al. 2013). An
object-based CD approach was designed in (Liu and Du 2010), which analyzes
different spectral and texture features extracted during the segmentation process from
two independent time images. A superpixel segmentation was applied to stacked
bitemporal images, and several derived features were used to describe changes in
the difference image according to the supervised classification (Wu et al. 2012).
An objected-based method was designed to create objects in each single-time image
according to segmentation and then generated different representative features (Wang
et al. 2018). A weighted Dempster–Shafer theory (wDST) fusion OBCD method was
proposed by combining multiple PBCD results, which can automatically calculate
and assign a certainty weight for each object of the PBCD result while considering
the stability of an object (Han et al. 2020). However, the selection of the optimal
segmentation scale is still an open issue and was mainly implemented based on the
empirical analysis in OBCD methods (Kaszta et al. 2016). Moreover, most of the
above existing work focused on solving a binary CD problem, and very few were
designed for dealing with the more challenging unsupervised multiclass CD case (Liu
et al. 2019b).
Accordingly, the following problems should be further analyzed: (1) how spatial
neighboring pixels have an impact on unsupervised change representation and
detection; (2) how to effectively model structural and geometric information of
change targets to enhance change representation; (3) how to integrate multiscale and
multidimensional change descriptions to increase change separability; (4) how to
adaptively find a suitable segmentation scale in the unsupervised OBCD approach.
In this chapter, we investigate these issues and develop new spectral–spatial CD
approaches in multitemporal multispectral images.
1.3. Unsupervised multiclass change detection approaches based on

modeling spectral–spatial information
1.3.1. Sequential spectral change vector analysis (S2 CVA)
Here, we first recall a popular state-of-the-art unsupervised multiclass CD method,

S2 CVA. The two proposed approaches are designed based on it. However, it is
important to note that the S2 CVA that is pixel-wise only relies on the analysis
of spectral information. Originally, from the standard C2 VA (Bovolo et al. 2012),
S2 CVA was specially proposed for multiclass CD in bitemporal hyperspectral images
according to a hierarchical analysis. It allows the visualization and detection of
multiple changes by considering all spectral channels, without neglecting any band. A
compressed 2D polar domain is built based on the construction of two change variables
(Liu et al. 2015), i.e. magnitude ρ and direction θ. More particularly, the magnitude ρ
is the Euclidean compression of SCVs. It measures the spectral brightness of changes.
The direction θ is built based on the spectral angle distance (SAD) (Keshava 2004),
which points out different changes with respect to the variances in spectral response
for a given pixel. However, note that the “from–to” class transition information is not
explained due to the unsupervised nature of S2 CVA. More specifically, the change
variables ρ and θ are defined as:
g g
fB fB
f ÿ ` ˘2 fÿ ` ˘
ρ“e Xb D “ e Xb ´ Xb 2
2 1 [1.1]
“
b 1 “
b 1
»¨ g ˛fi
fB
ÿ
B
` ˘ f ÿ ` ˘2 ÿ
B
θ “ arccos –˝ XD r {e
b b b
XD prb q2 ‚fl [1.2]
b 1“ b 1 “ “
b 1
b
where XD is the b-th (b=1,...,B) component of the differencing (SCVs) image XD ,
which was calculated on the co-registered images X1 and X2 , and B is the number
of image bands (i.e. the dimensionality of SCVs). rb is the b-th component of an
“ ?
adaptive ? ‰ r. In the standard C2 VA, r is2 defined as a unit constant vector
reference vector
r “ 1{ B, ..., 1{ B (Bovolo et al. 2012). In S CVA, it is improved as the first
eigenvector of the covariance matrix A of XD (Liu et al. 2015):
” ı
A “ cov pXD q “ E pXD ´ E rXD sq pXD ´ E rXD sqT [1.3]
where E rXD s is the expectation of XD . The eigen-decomposition of A can be

represented as:
A¨V“V¨W [1.4]
where W is a diagonal matrix with eigenvalues being sorted in a descending order

(i.e. λ1 ą λ2 ą ...λB ) in the diagonal, and V is the matrix of eigenvectors (i.e.
V “ rv1 , v2 , v3 , ..., vB s). The reference vector r is defined as the first eigenvector
v1 corresponding to the largest eigenvalue λ1 , i.e. r “ v1 . Note that this results in
an improved change representation by projecting the SCVs into a reference direction
that maximizes the variance of the measurement, while preserving the discriminative
information of different changes.
Figure 1.2. 2D polar change representation domain in the C2 VA method
Then a 2D compressed polar domain D is defined based on ρ and θ:
D “ tρ P r0, ρmax s and θ P r0, πsu [1.5]
where ρmax is the maximum value of ρ. A semicircle scattergram (see Figure 1.2) is
used for visualizing multiple changes in D. Regions of the semicircle SCn (which
represents the unchanged pixels) and the semiannulus SAc (which represents the
changed pixels) are mathematically defined as:
SCn “ pρ, θ : 0 ď ρ ă Tρ and 0 ď θ ď πq [1.6]
SAc “ pρ, θ : Tρ ď ρ ă ρmax and 0 ď θ ď πq [1.7]
Under the assumption that X1 and X2 are radiometrically and geometrically

corrected, in the 2D polar domain, the unchanged SCVs result in low magnitude
values close to zero and are thus distributed within SCn , whereas the changed SCVs
are represented in SAc having higher magnitudes. Different spectral behaviors of
the SCVs reflected on the direction variable, hence leading to the representation of
multiclass changes in each sector of SAc . The considered multiclass CD problem is
solved by defining a threshold Tρ along the magnitude ρ to separate the unchanged
and changed SCVs (i.e. associated with SCn and SAc , respectively), and to separate
multiple changes (C1 ,...,CK ) along the direction θ by setting multiple thresholds
Tθ,k (k=1,..., K-1) in SAc . Note that a hierarchical analysis is implemented to discover
and detect all possible changes (both strong and subtle changes) in hyperspectral
images due to complex change representations in the high dimensionality of
hyperspectral images. However, fewer levels may be obtained when only a few
spectral bands are considered, as in the multispectral case.
1.3.2. Multiscale morphological compressed change vector analysis
In this section, we introduce a proposed multiscale morphological compressed

change vector analysis (M2 C2 VA) method. It aims to investigate a proper way
to integrate multiscale spectral–spatial change information, especially to improve
multiclass change representation and discrimination in C2 VA. The proposed
approach consists of three main steps: 1) SCV reconstruction based on multiscale
morphological processing; 2) multiscale change information ensemble; 3) multiclass
change representation and discrimination. The block scheme of the proposed approach
is shown in Figure 1.3.
1.3.2.1. SCV reconstruction based on multiscale morphological processing

In the standard C2 VA, a SCV indicates a pixel that is unchanged or has a possible
kind of change according to its specific signature and constructed change variables
(i.e. ρ and θ). However, the original SCVs may contain abnormal spectral variations
and noises, which may lead to a high number of omission and commission errors. To
address this problem, the morphological profile (MP) is applied to better model and
preserve the geometrical structure of change targets. It is defined as a sequence of
mathematical closing and opening operations on the image with different structural
element (SE) sizes. In particular, opening by reconstruction (OR ) and closing by
reconstruction (CR ) (Benediktsson et al. 2005) for a gray-level image f are defined
as:
“ ‰
i
OR pf q “ Rδ εi pf q [1.8]
“ ‰
i
CR pf q “ Rε δi pf q [1.9]
where i is the radius of the SE. Here, δ i p¨q and εi p¨q are the dilation and erosion
operations, respectively. Rδ and Rε are the geodesic reconstruction by dilation and
erosion, respectively. In particular, the components of MP, i.e. OR and CR , are able
to suppress brighter and darker regions, respectively, that are smaller than the moving
SE, while preserving the geometrical characteristics of a region larger than the SE
(Dalla Mura et al. 2010). Small isolated objects are merged into a surrounding local
background while the main structure is kept. Due to the fact that different image
objects usually have different sizes, multiscale representation allows us to explore
different hypothetical spatial domains by using a range of SE sizes, in order to obtain
the best response for different structures (Mura et al. 2008).
12
SCVs Reconstructed SCVs

CD Map
X1
Multiscale Ensemble
Image Differencing
X
D Morphological Processing
based
Change Detection
X
2
Figure 1.3. Block scheme of the proposed M2 C2 VA technique

For the B-dimensional SCVs XD , at a given scale i, its OR and CR are also
B-dimensional:
` ˘ ` ˘ ` ˘(
i
OR pXD q “ ORi X1D , . . . , ORi XbD , . . . , ORi XBD [1.10]
` ˘ ` ˘ ` ˘(
i
CR pXD q “ CRi X1D , . . . , CRi XbD , . . . , CRi XBD [1.11]
b P r1, B s, i P r1, N s
Note that either OR i

pXD q or CRi pXD q can be used as an input for the detector
2
(e.g. S CVA), but ambiguities may arise due to the selection of a specific operator
(i.e. OR or CR ) and its consequential effects (i.e. suppression of brighter or darker
objects). The joint use of OR and CR is likely to be more reliable. In this work,
the four-connected neighborhood was considered, and the marker and mask image
represented the dilation result and the original input band (for OR ) (or complement
image of the original input band for CR ), respectively. A disk shape was selected
for the SE, which has been demonstrated to be a robust shape in different scenarios
(Benediktsson et al. 2005; Mura et al. 2008). The size of the SE i was increased from
1 to 6, in order to implement a multiscale analysis using the reconstructed SCVs.
1.3.2.2. Multiscale change information ensemble

By increasing the size of the SE, change objects can be modeled at different
scales, while exploring the interaction with the surrounding regions to preserve more
geometrical details. Accordingly, a more comprehensive description of change objects
is obtained through multiscale analysis. Therefore, a multiscale ensemble is conducted
on the reconstructed SCVs. Let OCR i
pXD q be the stacking of the reconstructed SCVs
(i.e. OR pXD q and CR pXD q) at a given size i, having a dimensionality of 2ˆB. It is
i i
defined as:
“ ‰
i
OCR pXD q “ ORi pXD q , CRi pXD q [1.12]
Then, the extended SCV S ru, v s is defined as an integration of sequential

i
OCR pXD q, with lower and upper bounds equal to u and v, respectively:
“ ‰
S ru, v s “ OCR
u
pXD q , OCRu`1 pXD q , . . . , OCRv pXD q [1.13]
Consequently, an extended SCV feature set with 2ˆB ˆM dimensionality is built,

where M is the length of components in the sequence, i.e. M “ v ´ u ` 1. It is worth
noting that S ru, v s extends the change representation along the spectral direction, as
well as considering the multiscale spatial information in the ensemble process. Then,
S ru, v s is used as the input for the detector.
14
SCVs Y′ CD Map
X1
X Superpixel-Level Spectral Decision

Image Differencing
D fusion based CD
Change Representation

X
2 Optimal Segmentation
Scale
Determination
Figure 1.4. Block scheme of the proposed superpixel-level multiclass CD approach

1.3.2.3. Multiclass change representation and discrimination

The aim of this step is to visualize and discriminate multiclass changes present in
the reconstructed SCVs. To this end, the S2 CVA detector introduced in section 1.3.1
is applied on the S ru, v s. Note that the detector exploits not only spectral variations,
but also spatial variations represented in the reconstructed SCV components. It is also
worth noting that the 2D polar scattergram projects multiclass change information
from the considered high-dimensional reconstructed SCVs into a low-dimensional
(i.e. 2D) feature space, which is lossy and ambiguous on the type of changes. However,
the most significant discriminative information of different changes is preserved.
Instead of using thresholding to segment the binary and multiple classes in the
variables ρ and θ, the simple but effective clustering, i.e. k-means, is used for
generating the final CD map, which does not rely on any specific data distribution.
This is due to the fact that changed and unchanged pixels in ρ and multiclass changes
in θ do not always follow a Gaussian mixture distribution (Zanetti et al. 2015). Thus,
for the binary CD step (i.e. separating two classes), the number of clusters kρ is equal
to 2, and for the multiclass CD step, the number kθ is defined as the number of changes
observed in the 2D polar scattergram.
1.3.3. Superpixel-level compressed change vector analysis
In this section, we propose an unsupervised superpixel-level compressed change

vector analysis (SPC2 VA) approach for multiclass CD. The traditional pixel-level
spectral change analysis is converted into the superpixel level. Therefore, the spectral
change representation and identification are regularized and enhanced under the
superpixel constraints. The block scheme of the proposed approach is shown in
Figure 1.4.
1.3.3.1. Superpixel-level spectral change representation

The original SCVs focus on the spectral variation representation from each
individual pixel, thus ignoring spatial correlation with neighboring pixels and the local
spectral homogeneity associated with real land-cover objects. This may lead to the
commission and omission errors, and a decrease in the overall detection accuracy.
Superpixel segmentation can capture image redundancy and generate convenient
primitives to compute representative features, while reducing the complexity of
the subsequent processing and analysis. In this work, we used the simple linear
iterative clustering (SLIC) algorithm (Achanta et al. 2012) as the core algorithm
for generating superpixel segments. Compared to the other popular segmentation
methods, SLIC offers a better performance in boundary adherence and generates
superpixels efficiently under the same hardware conditions. Moreover, SLIC is
memory efficient, and only requires the storage of the distance from each pixel to
its nearest cluster center. Most importantly, it can be smoothly integrated within
the proposed method, especially from the pixel to superpixel-level spectral change
representation and detection, which drives a proper algorithm utilization.
The general idea of the SLIC algorithm is to find small regional clusters by
considering their local homogeneity (Achanta et al. 2012). The key step is to calculate
the distance d that implements a measurement from the spectral–spatial point of view.
Let dcolor and dxy be the spectral and spatial distances between two given pixels α
and β, respectively, defined as:
b
dcolor “ pLα ´ Lβ q2 ` pAα ´ Aβ q2 ` pBα ´ Bβ q2 [1.14]
b
dxy “ pxα ´ xβ q2 ` pyα ´ yβ q2 [1.15]
Here, pL, A, B qT denotes the CIELAB color space values, with L being the color
lightness and A and B representing color values along red-green and blue-yellow
axes, respectively. px, y qT denotes the coordinates of a given pixel. A final weighted
distance measure dαβ can be defined as:
d ˆ ˙2
dα,β “ d2color ` m2
dxy
s
[1.16]
where s is the width of grids. It controls the size of created superpixels, i.e. the greater
the s,a
the larger the superpixels. A roughly equal-sized grid interval can be defined as
s “ pZ {N q, where Z is the total number of pixels and N is the desired number
1
of superpixels. In reality, the real number of generated superpixels (defined as N )
might be slightly different from N . The parameter m controls the relative importance
between the color similarity and the spatial proximity. The greater the m, the greater
the emphasis on spatial proximity and the compactness of a generated superpixel. A
regular m value can be defined within the range of [1, 40]. For more details, readers
can refer to the paper by Achanta et al. (2012).
In order to enhance the spectral variations due to the limited bands in multispectral
images, principal component analysis (PCA) is applied to the original SCVs. This
strengthens the change representation and extends the feature space. The first three
principal components (i.e. PCs) are used in the SLIC algorithm to generate the
segments (i.e. superpixels) with the identified boundaries. Then original SCVs are
stacked with the PCs to create an enhanced feature set (denoted as SCVs-PC). Note
that normalization is conducted on SCVs-PC bands to make data dynamic range
consistent. Finally, a mean operation is applied on each segment in the SCVs-PC
bands, where the mean vector is used to replace the original SCVs-PC vectors, in
order to achieve the enhanced spectral change representation at the superpixel level.
Note that, by doing this, the change information is concentrated with spectral–spatial
coherence and the computational cost is largely reduced when compared with the
original pixel-wise processing.
1.3.3.2. Determination of the optimal segmentation scale

As mentioned previously, the number of superpixels N and the compactness
factor m need to be determined in the SLIC algorithm. Note that in practical
implementations, the parameter m impacts less than N on the segmentation results.
Therefore, after multiple trials, we fixed m = 30 in this work. The only focus is
on the determination of the optimal segmentation scale parameter N . To this end,
an unsupervised strategy is applied based on the analysis of the global entropy.
Note that after the mean operation on each superpixel, the texture information in
the segments will be relatively suppressed, which may have an influence on the
following CD performance. The main idea of the used criterion is to evaluate the
information maintained in the superpixel-level segmented image inherited from the
original pixel-level image. Thus, the one-dimensional image entropy (Global Entropy,
GE) (Han et al. 2008) is calculated based on multi-scale segmentation results, where
the optimal segmentation scale is determined by analyzing the change of GE values:
ÿ
n
GE “´ pk log pk [1.17]
“
k 1
where n denotes the gray level and p “ ppk qk“1,2,...,n contains the histogram counts
1
of the first three bands of Y . It is worth noting that with the increasing N, GE values
are expected to increase continuously and approach the value of the original image.
The logarithmic function is then used to fit the GE results to estimate the threshold
for the optimal segmentation scale. A detailed description of this step is provided in
Table 1.1.
Step 1: Initialize N , which is approximated as the smaller one in the rows and columns of
the input image, and the segmentation scale interval is set approximately equal to (row +
column)/20.
Step 2: Calculate the GE value of each segmented image under different searching scales.
Step 3: Fit the logarithmic function on the obtained GE results and calculate the gradient.
Step 4: Estimate the optimal segmentation scale by analyzing the gradient of GE values,
where the convergence threshold TGE is defined approximately equal to 100/(row +
column) based on the input image.
Table 1.1. Determination of the optimal segmentation scale based on GE analysis

1.3.3.3. Decision fusion-based CD

The multiclass CD is implemented based on the change magnitude and direction
variables (i.e. ρ and θ) constructed by S2 CVA at the superpixel level. In paticular,
a decision fusion-based CD step is proposed to enhance and optimize the binary
CD output (see Figure 1.5), by analyzing the change magnitude of SCVs-PC bands.
Three binary CD results are taken into account based on a majority voting process,
to determine whether a given superpixel (segment) is changed or not. The first
input is made from the pixel-level binary CD result constrained by the segmented
boundaries, where a threshold Tρ is defined on ρ by using the EM algorithm under
a Bayesian framework (denoted as Bayesian-EM) (Bruzzone and Prieto 2000a).
Change and no-change pixels are counted within each superpixel, respectively, in
order to determine the label of that superpixel with respect to the counting majority.
The second and third inputs are generated at the superpixel level. The change
1
magnitude of superpixels is calculated on Y , and binary CD results are obtained
using Bayesian-EM thresholding and fuzzy c-means (FCM) clustering, respectively. A
three-input majority voting decision is applied by integrating three binary CD results
from a pixel-to-superpixel perspective. Note that the final binary CD decision is made
on each superpixel, depending on the majority label of three independent inputs, as
shown in Figure 1.5. FCM clustering is then applied on the variable θ but only focuses
on the changed superpixels, with the given estimated number of clusters equal to K in
the 2D compressed polar domain D, as shown in Figure 1.2.
1.4. Dataset description and experimental setup
1.4.1. Dataset description
The first dataset is a pair of QuickBird images acquired over an urban area in
Xuzhou city, Jiangsu province (China) in September 2004 (X1 ) and May 2005 (X2 ),
respectively. The Gram–Schmidt (G-S) pan-sharpening algorithm was applied to fuse
the multispectral (2.4 m) and panchromatic images (0.6 m), in order to generate
the final fused dataset (0.6 m) with four spectral bands (i.e. red, green, blue and
near infrared). Images were radiometrically corrected and co-registered (with residual
misregistration of about 0.3 pixels), and the final region with a size of 760ˆ370 pixels
was extracted for the CD experiment. A total of six change classes are presented in
this scenario. Detailed class transitions are: 1) soil to built-up area (C1); 2) soil to roof
(face to the sun) (C2); 3) soil to roof (back to the sun) (C3); 4) built-up area to soil
(C4); 5) shadow to soil (C5); 6) trees to soil (C6). The false color composite images
X1 and X2 are shown in Figure 1.6 (a) and (b), respectively, and Figure 1.6 (c) presents
the change reference map made by careful image interpretation.
Figure 1.5. Illustration of the decision-level fusion-based binary CD step.
For a color version of this figure, see www.iste.co.uk/atto/change1.zip
Unsupervised Change Detection in Multitemporal Remote Sensing Images
19
20
(a) (b) (c)

Figure 1.6. Color-infrared composite of bitemporal QuickBird pan-sharpened images (bands 4, 3, 2)

over an urban area in Xuzhou city, acquired in (a) 2004 (X1 ) and (b) 2005 (X2 ). (c) Change
reference map. For a color version of this figure, see www.iste.co.uk/atto/change1.zip
(a) (b) (c) (d) (e)
Figure 1.7. False color composite of QuickBird images of the Indonesia tsunami dataset acquired in: (a) April 2004 (before
tsunami), (b) January 2005 (after tsunami); (c) and (d) are two subsets selected from the whole scene in the qualitative
experimental analysis. (e) Change reference samples. For a color version of this figure, see www.iste.co.uk/atto/change1.zip
21
The second dataset is made up of a pair of co-registered QuickBird multispectral

images with a spatial resolution of 2.4 m, acquired over a large coastal area in
Indonesia in April 2004 and January 2005, respectively (see the two times of the
false color composite images shown in Figure 1.7(a) and (b), respectively). The whole
study area considered has a size of 3,250ˆ4,350 pixels. Two subsets were selected for
detailed qualitative analysis, as shown in Figure 1.7(c) and (d). A change reference
sample map was made by careful image interpretation (Figure 1.7(e)). The changes
that occurred in this scene were mainly caused by the tsunami in December 2004,
which include the following classes: (1) vegetation to flooded district (C1); (2) forest
to flooded district (C2); (3) shadow to vegetation (C3); (4) shadow to forest (C4);
(5) bareland to flooded district (C5); (6) cloud to forest (C6); (7) forest to shadow (C7).
1.4.2. Experimental setup
For comparison purposes, CD results obtained by the proposed M2 C2 VA and

SPC2 VA approaches were compared with two state-of-the-art unsupervised multiclass
CD techniques, including the iteratively reweighted multivariate alteration detection
(IR-MAD) (Nielsen 2007), and the sequential spectral change vector analysis
(S2 CVA) (Liu et al. 2015). Note that M2 C2 VA and SPC2 VA considered both spectral
and spatial change information, whereas IR-MAD and S2 CVA considered only the
spectral change information. Detailed quantitative and qualitative analyses were
conducted according to the obtained CD accuracy, i.e. OA and Kappa, and error
indices, i.e. omission errors (OE), commission errors (CE), total errors (TE), and
the obtained CD maps. In addition, the computational cost was also considered in
each method and compared. All of the experiments were conducted using MATLAB
R2016b, on an Intel (R) Core (TM) i7-6700 CPU @ 3.40GHz PC with 32 GB of
RAM.
Figure 1.8. 2D compressed change representation in the polar domain (Xuzhou

dataset). For a color version of this figure, see www.iste.co.uk/atto/change1.zip
(a) (b)
Figure 1.9. Determination of the optimal segmentation scale in the Xuzhou dataset.
(a) GE values of different segmented scales; (b) logarithmic fitting results based on
(a). For a color version of this figure, see www.iste.co.uk/atto/change1.zip
23
1.5. Results and discussion
1.5.1. Results on the Xuzhou dataset
The 2D compressed change representation in the polar domain is shown in

Figure 1.8, where six change clusters can be observed, indicating different change
directions. Then, K = 6 is fixed in the considered CD methods for performing the
multiclass CD.
The optimal segmentation scale was estimated by analyzing the GE values (see
Figure 1.9(a)) and its logarithmic fitting results (see Figure 1.9(b)). The TGE was
calculated and defined as 1 ˆ 10´4 , then the final segmentation optimal scale was
obtained as 1,160.
From the detailed quantitative and qualitative analyses based on the obtained CD
accuracy and error indices (see Table 1.2) and the obtained CD maps (see Figure 1.10),
it can be observed that the proposed M2 C2 VA and SPC2 VA approaches resulted in
higher performance than the reference methods, with respect to the higher OA and
Kappa values and smaller detection errors (see Table 1.2). In particular, SPC2 VA
achieved the highest accuracy (i.e. OA = 92.74% and Kappa = 0.7464), outperforming
the other methods. Improvement can also be observed from the obtained change maps
by comparing them with the reference map; the identified change targets are more
accurate in the two proposed approaches (i.e. Figure 1.10(c) and (d)) than those in the
two reference methods (i.e. Figure 1.10(a) and (b)). In two reference methods, S2 CVA
showed better performance than IR-MAD. As for the computational cost, the proposed
SPC2 VA approach exhibited efficient performance, which consumed less time than the
proposed M2 C2 VA and IR-MAD, and similar or slightly more time than S2 CVA (i.e.
8.91 s vs. 7.12 s), but yielding a significantly higher OA value (i.e. 92.74% vs. 87.74%).
Processing level CD methods OA Kappa CE OE TE Time cost

(Features) % (pixels) (pixels) (pixels) (s)
Pixel-level IR-MAD 85.23 0.5900 41,524 17,256 58,780 9.82
(Spectral) S2 CVA 87.74 0.6390 28,532 16,397 44,929 7.12
Pixel-level Proposed M2 C2 VA
91.86 0.7169 11,316 17,455 28,771 18.30
(Spectral–spatial) ([u, v] = [1, 6])
Superpixel-level) Proposed SPC2 VA
92.74 0.7464 8,673 16,411 25,084 8.91
(Spectral–spatial) (N = 1160, m = 30)
Table 1.2. Multiclass CD results obtained by the

proposed and reference methods (Xuzhou dataset)
1.5.2. Results on the Indonesia tsunami dataset
The 2D compressed change representation in the polar domain is shown in

Figure 1.11, where seven change clusters (i.e. K = 7) can be observed. The optimal
segmentation scale was estimated as 35,140 according to Figure 1.12. Note that the
CD experiment and the quantitative analysis were carried out on the whole image
scene.
(a) (b)
(c) (d)
Figure 1.10. Comparison of the multiclass CD maps obtained by: (a) IR-MAD;
(b) S2 CVA; (c) proposed M2 C2 VA ([u, v ] = [1, 6]); (d) proposed SPC2 VA with
N = 1,160 and m = 30 (Xuzhou dataset). For a color version of this figure, see
www.iste.co.uk/atto/change1.zip
Figure 1.11. 2D compressed change representation in the polar domain

(Indonesia tsunami dataset). For a color version of this figure,
see www.iste.co.uk/atto/change1.zip
26
(a) (b)
Figure 1.12. Determination of the optimal segmentation scale in the Indonesia tsunami dataset.
(a) GE values of different segmented scales; (b) logarithmic fitting results based on (a).
For a color version of this figure, see www.iste.co.uk/atto/change1.zip
Processing level CD methods OA Kappa CE OE TE Time cost

(Features) % (pixels) (pixels) (pixels) (s)
Pixel-level IR-MAD 81.09 0.4578 447,773 460,876 908,649 566.58
(Spectral) S2 CVA 86.04 0.6044 407,410 398,629 806,039 466.97
Pixel-level Proposed M2 C2 VA
91.15 0.7298 202,919 314,088 517,007 691.44
(Spectral–spatial) ([u, v] = [1, 6])
Superpixel-level Proposed SPC2 VA
93.69 0.8038 135,863 225,979 361,842 667.38
(Spectral–spatial) (N = 35140, m = 30)
Table 1.3. Multiclass CD results obtained by the proposed

and reference methods (Indonesia tsunami dataset)
We conducted a careful comparison between the qualitative and quantitative results

obtained by two proposed approaches (see the CD maps in Figure 1.13(c) and (d))
and two reference methods (see CD maps in Figure 1.13(a) and (b)) as in the previous
dataset. The quantitative evaluation was carried out based on the available change
reference samples, as shown in Figure 1.7(e); the numeric results are provided in
Table 1.3. Note that the number of multiclass changes was equal to 7 in all of the
methods, as estimated in the polar domain (see Figure 1.11).
From the CD maps shown at the global scale (first row) and the local scale (second
and third rows) in Figure 1.13 and the CD accuracies in Table 1.3, we can see that
the proposed approach SPC2 VA obtained the highest accuracy (i.e. OA: 93.69% and
Kappa: 0.8244) among all of the considered methods. The proposed M2 C2 VA also
showed good results and outperformed the two reference methods. Smaller detection
errors were found (i.e. 361,842 and 517,007 pixels in the two proposed approaches
against 908,649 and 806,039 pixels in the two reference methods), especially the
CE values. The detected change targets are more homogeneous and regular with
respect to their shapes and spatial distributions (see Figure 1.13(c) and (d)). This
demonstrates that the proposed spectral–spatial approaches are able to deal with
CD for large-scale data and offer a high detection performance, i.e. enhancing the
change targets and suppressing the no-change class. For two pixel-wise reference
methods, S2 CVA performed better than IR-MAD, as in the IR-MAD results, change
and no-change pixels are highly mixed in their spectral representation, leading to a
relatively poor detection accuracy. This is due to the complexity of the CD task in the
high spatial resolution multispectral images, and the pixel-wise correlation analysis
may fail to properly model the change targets solely based on spectral information.
From the point of view of the computational cost, the two proposed methods resulted
in a higher time cost than the two reference methods, but still at an acceptable level.
Therefore, from the careful qualitative analysis on both the global and local scales of
the obtained CD maps and the quantitative accuracy analysis, the effectiveness of the
proposed approaches in addressing a large complex multi-CD problem is validated.
(a) (b) (c) (d)

Figure 1.13. Comparison of the CD maps obtained by: (a) IR-MAD; (b) S2 CVA;
(c) proposed M2 C2 VA([u, v ] = [1, 6]; (d) proposed SPC2 VA with N = 35,140 and
m = 30, respectively, where the first row is the whole image scene, and the second
and third rows represent two subset results for a detailed visual comparison purpose
(Indonesia tsunami dataset). For a color version of this figure, see www.iste.co.uk/
atto/change1.zip
1.6. Conclusion
Owing to the automatic and unsupervised nature, unsupervised CD always

represents a very interesting and important CD research and application frontier.
However, the absence of ground reference and prior knowledge makes this practical
problem more challenging and complex than the supervised ones. In this chapter,
we reviewed the current development of unsupervised CD methods in multitemporal
multispectral remote sensing images, and analyzed existing open issues and
challenges. In particular, we focused on the spectral–spatial perspective to find robust
solutions to the important multiclass CD problem. Accordingly, two approaches were
proposed, including M2 C2 VA and SPC2 VA. By taking advantage of the spectral
and spatial joint analysis on the multispectral change representation, the original
pixel-level CD performance was enhanced by considering both the spectral variation at
the global scale and the spectral homogeneity and spatial connectivity and regularity of
change targets at the local scale. Experimental results obtained two real multispectral
datasets covering a complex urban scenario, and a large-scale tsunami disaster

scenario confirmed the effectiveness of the proposed approaches in terms of higher CD
accuracy and computational efficiency when compared with the reference methods.
For future works, advanced techniques still need to be designed to deal with more
complex real unsupervised CD cases, mainly focusing on, but not limited to the open
issues and challenges pointed out in section 1.2.2.
1.7. Acknowledgements
This work was supported by the Natural Science Foundation of China under Grant
42071324, 41601354, and by the Shanghai Rising-Star Program (21QA1409100).
1.8. References
Achanta, R., Shaji, A., Smith, K., Lucchi, A., Fua, P., Süsstrunk, S. (2012). Slic
superpixels compared to state-of-the-art superpixel methods. IEEE Transactions on
Pattern Analysis and Machine Intelligence, 34(11), 2274–2282.
Ban, Y. and Yousif, O. (2016). Change Detection Techniques: A Review. Springer
International Publishing, Cham.
Bazi, Y., Bruzzone, L., Melgani, F. (2005). An unsupervised approach based on the
generalized Gaussian model to automatic change detection in multitemporal SAR
images. IEEE Transactions on Geoscience and Remote Sensing, 43(4), 874–887.
Benediktsson, J.A., Palmason, J.A., Sveinsson, J.R. (2005). Classification of
hyperspectral data from urban areas based on extended morphological profiles.
IEEE Transactions on Geoscience and Remote Sensing, 43(3), 480–491.
Bouziani, M., Goïta, K., He, D.-C. (2010). Automatic change detection of buildings
in urban environment from very high spatial resolution images using existing
geodatabase and prior knowledge. ISPRS Journal of Photogrammetry and Remote
Sensing, 65(1), 143–153 [Online]. Available at: http://www.sciencedirect.com/
science/article/pii/S092427160900121X.
Bovolo, F. (2009). A multilevel parcel-based approach to change detection in very
high resolution multitemporal images. IEEE Geoscience and Remote Sensing
Letters, 6(1), 33–37.
Bovolo, F. and Bruzzone, L. (2007a). A split-based approach to unsupervised change
detection in large-size multitemporal images: Application to tsunami-damage
assessment. IEEE Transactions on Geoscience and Remote Sensing, 45(6),
1658–1670.
Bovolo, F. and Bruzzone, L. (2007b). A theoretical framework for unsupervised
change detection based on change vector analysis in the polar domain. IEEE
Transactions on Geoscience and Remote Sensing, 45(1), 218–236.
Bovolo, F. and Bruzzone, L. (2011). An adaptive thresholding approach to

multiple-change detection in multispectral images. IEEE International Geoscience
and Remote Sensing Symposium, 233–236.
Bovolo, F. and Bruzzone, L. (2015). The time variable in data fusion: A change
detection perspective. IEEE Geoscience and Remote Sensing Magazine, 3(3), 8–26.
Bovolo, F., Marchesi, S., Bruzzone, L. (2012). A framework for automatic and
unsupervised detection of multiple changes in multitemporal images. IEEE
Transactions on Geoscience and Remote Sensing, 50(6), 2196–2212.
Bruzzone, L. and Bovolo, F. (2013). A novel framework for the design
of change-detection systems for very-high-resolution remote sensing images.
Proceedings of the IEEE, 101(3), 609–630.
Bruzzone, L. and Prieto, D.F. (2000a). Automatic analysis of the difference image
for unsupervised change detection. IEEE Transactions on Geoscience and Remote
Sensing, 38(3), 1171–1182.
Bruzzone, L. and Prieto, D.F. (2000b). A minimum-cost thresholding technique for
unsupervised change detection. International Journal of Remote Sensing, 21(18),
3539–3544 [Online]. Available at: https://doi.org/10.1080/014311600750037552.
Celik, T. (2009). Unsupervised change detection in satellite images using principal
component analysis and k-means clustering. IEEE Geoscience and Remote Sensing
Letters, 6(4), 772–776.
Celik, T. and Ma, K.K. (2011). Multitemporal image change detection using
undecimated discrete wavelet transform and active contours. IEEE Transactions on
Geoscience and Remote Sensing, 49(2), 706–716.
Chen, J., Gong, P., He, C., Pu, R., Shi, P. (2003). Land-use/land-cover change
detection using improved change-vector analysis. Photogrammetric Engineering
and Remote Sensing, 69(4), 369–379.
Chen, G., Hay, G.J., Carvalho, L.M.T., Wulder, M.A. (2012). Object-based change
detection. International Journal of Remote Sensing, 33(14), 4434–4457 [Online].
Available at: https://doi.org/10.1080/01431161.2011.648285.
Coppin, P., Jonckheere, I., Nackaerts, K., Muys, B., Lambin, E. (2004). Review
article digital change detection methods in ecosystem monitoring: A review.
International Journal of Remote Sensing, 25(9), 1565–1596 [Online]. Available at:
https://doi.org/10.1080/0143116031000101675.
Dalla Mura, M., Benediktsson, J.A., Waske, B., Bruzzone, L. (2010). Morphological
attribute profiles for the analysis of very high resolution images. IEEE Transactions
on Geoscience and Remote Sensing, 48(10), 3747–3762.
Du, P., Liu, S., Gamba, P., Tan, K., Xia, J. (2012). Fusion of difference images for
change detection over urban areas. IEEE Journal of Selected Topics in Applied
Earth Observations and Remote Sensing, 5(4), 1076–1086.
Du, P., Liu, S., Xia, J., Zhao, Y. (2013). Information fusion techniques for
change detection from multi-temporal remote sensing images. Information Fusion,
14(1), 19–27 [Online]. Available at: http://www.sciencedirect.com/science/article/
pii/S1566253512000565.
Falco, N., Mura, M.D., Bovolo, F., Benediktsson, J.A., Bruzzone, L. (2013).
Change detection in VHR images based on morphological attribute profiles. IEEE
Geoscience and Remote Sensing Letters, 10(3), 636–640.
Ghosh, A., Mishra, N.S., Ghosh, S. (2011). Fuzzy clustering algorithms for
unsupervised change detection in remote sensing images. Information Sciences,
181(4), 699–715 [Online]. Available at: http://www.sciencedirect.com/science/
article/pii/S0020025510005153.
Han, P., Gong, J., Li, Z. (2008). A new approach for choice of optimal spatial scale
in image classification based on entropy. Wuhan Daxue Xuebao (Xinxi Kexue Ban)/
Geomatics and Information Science of Wuhan University, 033(7), 676–679.
Han, Y., Javed, A., Jung, S., Liu, S. (2020). Object-based change detection of
very high resolution images by fusing pixel-based change detection results using
weighted Dempster–Shafer theory. Remote Sensing, 12(6) [Online]. Available at:
https://www.mdpi.com/2072-4292/12/6/983.
Huang, X., Zhang, L., Zhu, T. (2014). Building change detection from multitemporal
high-resolution remotely sensed images based on a morphological building index.
IEEE Journal of Selected Topics in Applied Earth Observations and Remote
Sensing, 7(1), 105–115.
Hussain, M., Chen, D., Cheng, A., Wei, H., Stanley, D. (2013). Change detection
from remotely sensed images: From pixel-based to object-based approaches. ISPRS
Journal of Photogrammetry and Remote Sensing, 80, 91–106 [Online]. Available at:
http://www.sciencedirect.com/science/article/pii/S0924271613000804.
Kaszta, A., Van De Kerchove, R., Ramoelo, A., Cho, M.A., Madonsela, S., Mathieu, R.,
Wolff, E. (2016). Seasonal separation of African savanna components using
WorldView-2 imagery: A comparison of pixel- and object-based approaches and
selected classification algorithms. Remote Sensing, 8(9) [Online]. Available at:
https://www.mdpi.com/2072-4292/8/9/763.
Keshava, N. (2004). Distance metrics and band selection in hyperspectral processing
with applications to material identification and spectral libraries. IEEE Transactions
on Geoscience and Remote Sensing, 42(7), 1552–1565.
Khan, S.H., He, X., Porikli, F., Bennamoun, M. (2017). Forest change detection
in incomplete satellite images with deep neural networks. IEEE Transactions on
Geoscience and Remote Sensing, 55(9), 5407–5423.
Leichtle, T., Geiß, C., Wurm, M., Lakes, T., Taubenböck, H. (2017). Unsupervised
change detection in VHR remote sensing imagery – An object-based clustering
approach in a dynamic urban environment. International Journal of Applied
Earth Observation and Geoinformation, 54, 15–27 [Online]. Available at:
http://www.sciencedirect.com/science/article/pii/S0303243416301490.
Li, H., Celik, T., Longbotham, N., Emery, W.J. (2015). Gabor feature based
unsupervised change detection of multitemporal SAR images based on two-level
clustering. IEEE Geoscience and Remote Sensing Letters, 12(12), 2458–2462.
Liu, S. and Du, P. (2010). Object-oriented change detection from multi-temporal
remotely sensed images. Geographic Object-Based Image Analysis, number
XXXVIII-4/C7.
Liu, S., Bruzzone, L., Bovolo, F., Du, P. (2012). Unsupervised hierarchical
spectral analysis for change detection in hyperspectral images. 4th Workshop
on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing
(WHISPERS), pp. 1–4.
Liu, S., Bruzzone, L., Bovolo, F., Zanetti, M., Du, P. (2015). Sequential spectral
change vector analysis for iteratively discovering and detecting multiple changes
in hyperspectral images. IEEE Transactions on Geoscience and Remote Sensing,
53(8), 4363–4378.
Liu, S., Du, Q., Tong, X., Samat, A., Bruzzone, L., Bovolo, F. (2017a). Multiscale
morphological compressed change vector analysis for unsupervised multiple
change detection. IEEE Journal of Selected Topics in Applied Earth Observations
and Remote Sensing, 10(9), 4124–4137.
Liu, S., Tong, X., Bruzzone, L., Du, P. (2017b). A novel semisupervised framework for
multiple change detection in hyperspectral images. IEEE International Geoscience
and Remote Sensing Symposium (IGARSS), 173–176.
Liu, J., Chen, K., Xu, G., Li, H., Yan, M., Diao, W., Sun, X. (2019a). Semi-supervised
change detection based on graphs with generative adversarial networks. IEEE
International Geoscience and Remote Sensing Symposium (IGARSS), 74–77.
Liu, S., Du, Q., Tong, X., Samat, A., Bruzzone, L. (2019b). Unsupervised
change detection in multispectral remote sensing images via spectral-spatial band
expansion. IEEE Journal of Selected Topics in Applied Earth Observations and
Remote Sensing, 12(9), 3578–3587.
Another random document with
no related content on Scribd:
these do more harm than good, and spoil young people for ordinary
life. Whatever may be said concerning the dull hotch-potch turned
out “in bulk” by the American film industry, nobody can deny that an
element of romance has been brought into the lives of countless
poor people by Edison’s and Berliner’s great invention.
As regards wireless broadcasting, it is too early to speak of its
permanent effects. It has come upon us like an avalanche. For the
first time, it has enlisted youth in the highest electrical problems. It
has paved the way for the general diffusion of scientific knowledge
while ostensibly popularizing the art of music. In any case, it is a
complete answer to Ben Akiba and his saying about there being
nothing new.
Mr. J. B. S. Haldane[2] believes that the centre of scientific interest
now lies in biology, and that physiology will eventually invade and
destroy mathematical physics. It is quite possible that the advance of
what physicists sometimes playfully call the “inexact” sciences may
cast those of physics and chemistry entirely in the shade, but
although biology has made some difference to human life in the last
generation, its effects cannot remotely compare with those of
physical and chemical discoveries. The mere increase of speed in
transport, due to the internal-combustion engine, has caused a
speeding-up of the whole nervous system and a brightening of the
intelligence of all but an insignificant fraction of the population. When
the choice lies between the Quick and the Dead, even the most
sluggish temperament will put on a spurt, and this continued
sprinting across the motor traffic has produced a more agile
generation. If such a profound difference can be made in twenty
years, what will be the effect of even 200 years of continually
accelerated progress?
[2] Dædalus, pp. 10 and 16.
The continual acceleration of the rate of progress must not be lost
sight of in forecasting the future. It happens to coincide with a similar
increase in the consumption of accumulated fuels, like coal and oil,
and might be expected to slow down when those supplies of
preserved sunlight come to an end. But by that time other
accumulations will no doubt have been discovered and utilized.
Unless the Russian blight extends over Europe and America, we
may confidently look forward to a long era of steadily accelerated
progress. What form that progress will take is notoriously difficult to
forecast. The main difficulty arises from the fact that the most
promising discoveries sometimes turn out to be impracticable, or at
least of quite secondary importance. Nobody prophesied the great
development of the motor car, nor of that gigantic child of the old
Zoetrope or “Wheel of Life” which we call the Picture Theatre. A
genius like H. G. Wells could indeed write a marvellously accurate
forecast of flying achievements, but in one of his books he makes a
great deal of the Brennan mono-rail, which, after a sensational
beginning, failed to reach maturity, probably owing to the temporary
failure of that much-maligned but quite essential fertilizer of
inventions, capital. Other inventions, such as the speaking film, bear
within themselves certain weaknesses which may prevent them from
attaining great popularity.
But previous failures to peep into the future effectively shall not
deter the author from another attempt to pierce the veil hiding our
fate. He will proceed by “extrapolating” the curves representing
progress made hitherto, but will always allow for unexpected new
departures in what are already known to be possible directions.
Transport and Communications.—The most conspicuous changes

in our mode of life have been brought about by improved means of
transport. Railways and steamships have become essential to Great
Britain, largely owing to the fact that food-supplies have to be
purchased in exchange for exported manufactures. But even in non-
industrial countries the railway has entered very largely into the life
of the people. It is part of the essential attribute of life which seeks
diversity and a fuller activity of the senses. Life is, in effect,
prolonged when it is made to contain more and more detail. The
social life of a person is roughly measured by the number of people
with whom he converses in the course of a day. Improved transport
means facilities for extending this number. It also means an
enlargement of a person’s sphere of influence, of his “area of
effective occupation.” In business, it gives an opportunity of
increasing turnover, and thus reducing costs. It also facilitates
competition, both in buying and selling, and thus tends to enforce the
law of supply and demand and other enactments of what used to be
called the “dismal science,” but now ranks as the most “actual” of all
the sciences.
Postal and telegraphic facilities have the same general effect of
diversifying and extending life. The telephone has produced an
entirely new form of social intercourse, and a new privileged class.
By acquiring a telephone number, we obtain entrance into a hall
where rapid and varied communication becomes an abundant
source of activity, information, and amusement.
Increased facilities of this kind also tend to cement a nation of one
language into a closely organized whole, so that its cohesion and its
influence abroad are strengthened. The “temperature” of the national
life is raised, and it increases in proportion. An increased energy also
accelerates the rate of progress, so that it tends to spread like a fire.
What this acceleration will mean we can as yet only dimly surmise. A
time will come, no doubt, when we may crowd into an hour a variety
of experiences which our ancestors would have extended over a
lifetime. We need not necessarily do so, but the mere fact that it is
possible will add a zest and a richness to life such as we, with all our
advantages, can as yet hardly conceive. But the general effect will
be to reduce more and more the limitations now imposed upon us by
space. It will no doubt be eventually possible to get into
communication with anybody on earth at a moment’s notice,
provided that person is willing.
Privacy.—The last proviso is important. Civilization not only makes
us more accessible to those we appreciate, but also makes us less
accessible to those of whom we disapprove. An increase in our area
of choice would be of little value if everybody else could choose to
intrude upon us at any time or place. A limitation of social intercourse
to a chosen few, or its total cessation for the time required for rest
and recuperation in a strenuous life is one of the greatest boons one
can desire. The rank and file is more gregarious than the élite and,
as one of the ideals of progress is to raise the masses towards the
level occupied by the élite, the extension of facilities for seclusion is
of the essence of progress. Hence commons and open spaces are
provided in and around well-planned cities. Houses are provided with
sound-proof walls, and gardens are protected as much as possible
from “overlooking.” A very modern problem of the same kind is the
protection of wireless listeners from oscillations produced in
neighbouring receiving sets.
Clothing.—The present generation of civilized humanity justly
prides itself on its sensible style of clothing. Although masculine
clothing contains a number of “vestigial” elements which are absurd
survivals of former necessities, the practice, observed more
especially in Great Britain, of wearing a variety of apparel suited to
special occasions and occupations adds a certain elasticity to
fashions which otherwise are almost comically rigid. If a man wore a
wreath of flowers round his head instead of a hat he would not walk
fifty yards in any London street without being arrested for “insulting
behaviour” or “conduct likely to cause a breach of the peace.” Yet a
woman could do so with impunity if she did it with the necessary air
of assurance.
The future of clothing is largely affected by social problems.
Masculine clothing ceased to be demonstrative in Europe shortly
after the French Revolution. While aristocracy flourished, it paraded
itself aggressively in velvet and ruffles and powdered wigs. In
modern times it is bad form to strut and swagger, except on very
special occasions. Good clothing is not aggressively ornate or
expensive. Its quality is a matter of lines and cut and finish. Its
preciousness is disguised from the uninitiated. It is like the sober and
almost dingy town houses of the nobility, whose splendour is only
shown to favoured friends and trusted servants. In 1794 the dress of
the “aristo” was a passport to the guillotine. In later and wiser days it
is a disguise to deceive the tax-collector and the demagogue, and to
lull their prying rapacity into inactivity.
Feminine attire follows the same principles, modified by the
essential differences between the sexes. The ideal of a well-dressed
woman varies widely with her surroundings. In bad weather out of
doors, or in an unsympathetic crowd, her garments will be a
defensive armour designed to reveal as little of her personality as is
compatible with her purpose. Under more favourable conditions, they
will become a setting made as suitable as possible to the peculiar
qualities and attractions of the jewel they are supposed to contain.
Not every woman can, under present conditions, be beautiful, but
she can suggest beauty at every turn, remind us of beautiful things,
and give us that feeling of holy calm which we experience in the
presence of beauty, if she will but dress appropriately to the occasion
and to her own personality.
These considerations must affect our view of the appropriate
styles of a.d. 2025. Dress appropriate to the occasion! Dress is, after
all, a sort of extension of the physical personality. The body of a
naked child at play is the most perfect thing in beauty that can be
seen. All its muscles adapt themselves instantly to its activities.
Everything is appropriate and harmonious. A thin and clinging
covering would detract but little from its grace of movement and
expression, and might, indeed, add something of force and swiftness
that cannot be perceived in the mere play of muscles. Thus the
plumes of a Red Indian add to the sense of speed and purpose
conveyed by his movements.
I do not agree with Mr. H. G. Wells that the final ideal of clothing is
its total abolition. Clothing has the effect of enlarging man’s sphere
of activity until it covers the entire globe from the poles to the
equator. Another advantage is that it emphasizes mental qualities
rather than physical qualities. If mere physical beauty were the one
essential to human well-being, mankind would have long ago
insisted on its being freely displayed—and judged—without the
disguise of clothing.
Instead, a common agreement among civilized peoples insists that
on everyday occasions little but the face is to be visible, because its
features and expression give a clue to the mentality behind them. On
special occasions, such as balls and dinner parties, more may be
revealed by the gentler sex, but even then the area revealed must be
confined to what is least likely to show defects and is of least
physiological interest and importance, so that the attention may still
be directed towards mental rather than emotional or physical
qualities.
The evolution of clothing will, therefore, be in the direction of
adaptability to climate and occupation. New fabrics will no doubt be
invented, combining the warmth of fur with the softness and flexibility
of silk and the strength of linen. Dress will be light, so that half a
dozen changes of costume can be carried in a handbag, and will be
so designed that each change will involve no more inconvenience
than does the removal of a raincoat. And so we shall eventually
combine the Greek ideal of expressive drapery with the exacting
conditions of a strenuous modern life.
Housing.—Man is an animal with a cubical shell. If the earth were
reduced to the size of a football and its surface were examined with
a powerful microscope, we should see it studded with incrustations
like dried salt, especially about the river mouths. These incrustations
would be the cities, consisting of thousands of rectangular or cube-
shaped blocks. With a greater magnification we might see minute
specks swarming about these shell-like houses, elongated specks
with their longer axis vertical, and with a marked tendency to enter
the houses at nightfall and emerge again after daybreak.
If a giant had been watching the development of these
incrustations for several thousand years, he would have seen them
spreading from the Mediterranean and some parts of Asia till they
studded the Atlantic regions. Thence, after a time, they would spread
to the other side of the Atlantic, and become particularly numerous
along its western shores. Watching them again for several thousand
years, he might see these incrustations gradually dissolved, and the
disease—he would probably call it a disease—become “generalized”
all over the planet, the local incrustations giving way before a
universal but only slightly crusted condition of the earth’s entire land
surface.
This is the most probable solution of the general housing problem.
Ordinary and wireless telephony, soon to be supplemented by
“television,” will gradually reduce the isolation brought about by mere
space, while underground and overground transport of goods will
render the distribution of supplies less and less laborious. Houses
will, therefore, be built more widely apart than they are in cities, and
each will have its own private grounds. The structures will be of a
material impervious to heat and cold, but transparent or translucent
to light, though there will be means of darkening the whole house if
desired. Artificial lighting will not be by lamps, but by a close imitation
of diffused daylight, which is coolest and most restful to the eyes.
There will be no domestic servants. All the “work” of the house will
be done by machinery requiring but the turning of a switch and the
aiming of implements resembling magic wands. Cooking will be a
pleasant domestic art, most of the preparations being made by the
purveyors of food stuffs. As it is unlikely that the anatomical structure
and the physiological functioning of the human frame will be
materially changed for thousands of years to come, food will not be
very different from what it is now, but there will be a nicer
discrimination of what foods, and what quantity of them, are best
calculated to maintain perfect health.
Children.—Most prophets concerning the future of the human race
postulate many and radical changes in the birth and rearing of
children. Some say that advancing civilization will make the present
process impossible on account of the steadily increasing size of the
human skull, which will eventually make normal birth impossible.
Professor Haldane forecasts “ectogenesis” or the artificial ripening of
the embryo outside the human body. Whatever may happen to the
physical act of birth—it will no doubt be alleviated in many ways—
one hopes that the tender joys of watching over the development of
a child’s body and mind will not be taken away from us.
Much more enlightened care will, no doubt, be bestowed upon the
welfare of the infant than is done at present. How many crimes are
unwittingly committed against a child’s mentality by ignorant parents
and nurses! Lies and prevarications and evasions are always
reprehensible, but with children they are of fatal and life-long effect.
We owe the truth to a child more than to any adult. Our promises to
a child should be as binding as an oath. Tell them fairy tales by all
means, but tell them with a voice and expression which inevitably
stamps them as such, and makes belief optional.
I cannot see any effective substitute for family life so long as there
are children to bring up. Children feed on love as they do on food
and fresh air, and no vicarious love can take the place of the natural
affection between children and their parents. The institution of
marriage may undergo many and far-reaching changes[3], but
nothing is likely to change the paramount necessity of parental care
of, and responsibility for, children. Children are rooted in their
parents. They are, in a sense, survivals of their parents’ personality,
and constitute their chance of physical immortality. It is, therefore,
absurd to suppose that the human race will at any time in its history
consent to the “nationalization” of its children. On the other hand, the
supervision of ill-disposed or incompetent parents by the State will,
no doubt, become more and more strict.
[3] Its early history shows it to be a contrivance for the safe-
guarding of infants in their earliest years by allocating a large
proportion of responsibility to the father. Among Semitic and other
Asiatic tribes, this allocation of responsibility was fortified by the
somewhat crude precaution of secluding women. Marriage as a
means of allocating responsibility will become superfluous as
soon as parentage can be infallibly traced with the aid of the
microscope, as some biologists confidently foretell. The chief
raison d’etre of marriage will then be gone, but it will no doubt
continue for a considerable time as a picturesque survival of an
ancient custom.
Education.—In spite of the prodigious advance in educational
methods in the last two generations, education is still in a state of
primitive barbarity. We may, therefore, expect some very profound
changes in the centuries to come. There is still too much of the
methods of the pump about our education. The idea seems to be
that the teacher draws from the well of knowledge and administers
copious draughts to his pupils, and when they have swallowed these
they are educated! There is no better illustration of this curious view
than the modern method of imparting “higher” education. The
University lecture is, of all methods of imparting knowledge, about
the least effective. The student sits in a stiff attitude and maintains a
pose of strained attention. He endeavours to keep his mind
concentrated on the words and meaning of the professor. Every now
and then he succeeds, but then his thoughts persist in following their
own train of associations and the thread is broken. He jots down
disconnected notes, hoping to piece them together afterwards. This
piecing together is often the only process which really advances his
knowledge. It brings his own will-power and faculties into action. The
lecture only requires will-power for concentration on somebody
else’s thought, and this effort is negative and sterile.
If lectures must be, then they should be interrupted after every ten
minutes or so. The lecturer should then sit down and invite and
encourage his students to ask pertinent questions or advance sound
criticism.
In a class-room it is easier to keep the interest of the pupils alive.
Every effort should be made to let the information come from the
pupils rather than the teacher. In teaching history, for instance, I
should not have set lessons at all, but ask the pupils to collect facts
within a certain period, and reward them in accordance with their
success in presenting the facts and linking them up with others.
The education of the future will be like the medicine of the future.
Both will aim at eliciting and enlisting the powers of the pupil (or
patient) rather than dosing them. For the real learning and the real
cure must come from them.
Every normal child is anxious to learn, and can be easily brought
to feel and appreciate the intellectual joy of comprehension. But in
most children this joy is marred in early infancy by insufficient
attention to their struggles to understand the great world about them.
It is the years of infancy—the pre-school years—which are most
important in forming habits of thought. The closest watch should be
kept for early efforts at trains of reasoning. These efforts begin at the
age of three or thereabouts. They are often absurd and ludicrous,
but they should be treated with an indulgent and helpful respect, and
wrong conclusions should be modified, not by contradiction, but by
conviction of the contrary by example. If that is done, the child will
learn to trust his own powers of reasoning. If it is neglected, the
child’s mind will become shallow and unenterprising.
No child that can talk is too young to be asked for his opinion. He
will enjoy stating it, and will, as a rule, receive protests or contrary
opinions with interest and amusement.
All this may be a “counsel of perfection” to parents who are too
busy to look after their infants themselves and are content to entrust
their tender minds to more or less incompetent nurses. But the future
will realize more and more the great importance of the growing
minds of infants. In the United States this is to a large extent the
case already, and, as a consequence, their infants are the brightest
and most delightful creatures imaginable.
Labour.—In one of his earliest works, “The Time Machine,” Mr. H.
G. Wells forecasts a development of the labour situation very
different from that of the ordinary socialist Utopia. He figures an
arrangement by which all labour is done underground, and is done
by creatures (one can hardly call them “people”) whose bodies and
minds are thoroughly adapted to their task. The picture seems to be
a skit on the Victorian idea of the “family” upstairs and the servants
in the basement, but a grim and novel touch is added by the
information that the workers actually live on those that dwell in the
light, coming up at night to take them away in their sleep.
Such a solution, though it may draw some justification from the
bee-hive or the ant-heap, is not at all likely to be adopted by the
human race of the future. The essential service of Christianity, the
kernel which will remain after the mythological and dogmatic
accretions have been shed, is to provide mankind with an
imagination capable of conceiving and realizing the sufferings of
other people and creatures, and the will to remedy or obviate them
as far as possible. Through centuries of abuse, neglect and
misinterpretation that gift has gradually worked into the mass of
civilized humanity. The humanitarian ideal is explicit and articulate in
France, while in England it is disguised under such expressions as
“decency” and “playing the game,” or “live and let live.”
That it has not yet remade the world is due mainly to two causes:
the arrogance of those to whom money or social position gives an
advantage over smaller people, and the hatred and mistrust
engendered by this arrogance among the masses.
A workman in a physically fit condition does not object to working.
He sometimes feels the drudgery, boredom, or discomfort of it, but if
he is a skilled craftsman, his pride in his work gives him an interest
and satisfaction which helps him over many hours of toil. What he
does object to is to be driven and bullied by an unsympathetic and
perhaps unjust overseer, who turns out the master on every
occasion and lets him feel his power. He wants and sometimes
admires a leader, but he does not want a slave-driver.
In war, there is the same difference between the “come on” officer
and the “go-on” officer, and the same effect on discipline.
I believe that the co-operation of larger organized masses of men
will not only be required in the future as in the past, but that its scale
will eventually exceed anything yet seen in our history.
The key to the smooth working of such organizations is the spirit in
which the enterprise is undertaken. Let us give two examples, one
from the present day and one from, say, the year 2,000.
(1) a.d. 1925:

(a) Navvies required. Apply Eastern Counties
Railway Co.’s Depot....
(b) Eastern Counties Railway Co., Ltd. The 250,000
5% Cumulative Participating Preference Shares
are now offered for public subscription, payable
as follows....
(2) a.d. 2000:

Norfolk Water Supply Undertaking. A public meeting
will be held on ... at the Norwich Auditorium, to
announce and explain the purpose and plan of
this Undertaking. Workers and Contributors will
be enlisted on terms to be announced at this
meeting. Qualification papers may be sent to
the undersigned....
It is obvious that the “contributor” who places his savings at the

disposal of the Undertaking must not only not lose them thereby, but
must be compensated according to the risk he runs. In effect, he
provides the manual worker with the food he requires to exert his
strength, and he does so without any immediate benefit to himself. If,
after the Undertaking had got to work, his “contribution” were simply
refunded, he would receive no reward for his public service, a
service which implied a reduction in his own resources. This would
be unjust and would make his treatment worse than that of the
wage-earner.
To this simple plea a Communist would rejoin that nobody should
be allowed to have savings or accumulated resources or private
property of any kind. The absurdity of this contention is very easily
demonstrated, but let us add a few hard cases to the usual
arguments:—
(1) A tribe of shepherds requires water. There is one man who
knows the location of a well. His knowledge is his private property,
which he offers to sell for a number of cows. What should be done, if
the man is (a) accessible and undefended; (b) inaccessible.
(2) A girl has a beautiful head of hair, which could be sold abroad
for a considerable sum. Whose property is it?
(3) In a small and isolated town, 20 per cent. of the inhabitants
have good sets of false teeth. Another 20 per cent. require false
teeth, but have no means of getting them. What will be the action
taken by the Communist municipality?
The workman going to his work has quite a respectable capital to

accompany him on his way. He has clothes to keep him warm, boots
to save his feet from wearing out, a set of tools, perhaps, and
accumulated stores of food in every muscle of his body. Even if his
stomach is empty, he is still capable of work, although it will probably
be of inferior quality. There was a craze towards the end of last
century for living without food, the idea being that the body was not
nourished by food but by “vibrations” of some mysterious kind. The
craze, for obvious reasons, did not last long, but some prodigies of
fasting were performed while it lasted, and the leaders of the
movement prided themselves on doing their daily work and business
as usual. The truth was that, like the badger, they were living on their
own fat and consuming their own tissues. In the end they were
thinner and sadder, but little the worse otherwise.
Everybody in a physically fit condition is necessarily a capitalist.
That arch-capitalist among the animals, the squirrel, jumps on the
shoulders of the amiable Communists strolling in Regent’s Park and
then goes to “rattle in his hoard of acorns.”
There will be no discouragement of the acquisition of private
property in the centuries to come, but care will be taken that the
happiness accruing to the owner is not set off by misery inflicted on
others. Misery due to mere envy or jealousy will not be considered,
but, as a matter of fact, there is much less of that than there is
commonly supposed. The spectators of the Lord Mayor’s Show do
not envy the Lord Mayor his pomp and magnificence, but they are
there to delight their eyes with an unusual display. The crowds
collecting at the church door to see the blushing bride do not come
to turn green with envy, but to feast their eyes on something
exquisite and heartening to behold.
And the work of the future, the labour of thousands on a great
enterprise, will be accomplished in the spirit of adventure and
comradeship, like an Arctic expedition, let us say, and the distribution
of rewards will be conducted in the same spirit as the prize-giving at
a sports meeting.
The spirit of adventure is a most valuable incentive to work, an
incentive that is much neglected at present. It is part of the
workman’s grievance against capital that the capitalist has the
adventurous part of the undertaking.
The need for adventure in everyday life is proved by the enormous
prevalence of betting on horses, which is about the most stupid way
of adventuring money that one can conceive. The prodigies
performed under some piece-work agreements show that it is not the
quantity of work that is felt as a burden, but the consciousness of
being driven by a will other than one’s own. Many schemes of profit-
sharing also introduce an element of chance which is most welcome
to the worker, though logically it should be under a proviso of loss-
sharing in the same proportion.
A profit-and-loss-sharing understanding between employers and
workers would make every worker a capitalist. Indeed, as soon as
the workers (or the State, it matters not which) are sufficiently
organized to guard workers from undeserved destitution, it may be
quite feasible to organize public undertakings without wages of any
kind. Each worker would contribute “capital” in the form of a certain
amount of work. In case of a total loss on the undertaking, he would
receive no reward or wages whatever, any more than the capitalist
who engages in a profitless scheme. If the undertaking succeeded,
he would have a permanent interest in the revenue from it, in
proportion to the work contributed. Since it is easier to assess the
value of manual piece-work than mental work, it might well happen
that, though no wages were paid, salaries would be paid to
organizers, architects, overseers, and the like, who would thus be
the only “proletarians” in the concern!
Government.—“The first duty of a Government is to govern.” This
platitude sounds as if there were some hidden pearl of meaning
behind it. But on etymological analysis we find that the duty of
Government is to work the rudder of the ship of State. It is to give a
general direction to the activities of its citizens. The same or a similar
word is used in all countries based upon Roman law and citizenship.
The German equivalent, Regierung, is different. It means reigning or
doing the business of royalty, but as the origin of the word “Rex” is
the same as that of the main syllable of the words “Rector” and
“Director,” it comes to the same thing, a “directing” action.
Will this directing or controlling action ever become superfluous?
Its chief function is at present to determine the policy of a large
aggregate of human beings. In democratic countries this
determination is settled by a majority vote, from which there is no
legal appeal, though passive resistance or the threat of rebellion are
weapons which a minority can sometimes use with effect. A majority
vote would be practically certain to be wrong on most questions of
the day, but fortunately such a vote has no effect unless it is put into
action by a body of expert politicians comprising the Ministry, and
these again are largely guided by experts in the particular matter
with which the vote in question is concerned. Laws are not made by
Parliament. They are born in Government offices under expert
advice. The Cabinet, on the advice of its experts, decides to bring in
a Bill which is likely to be of some benefit, will probably be passed by
Parliament, and will encounter no serious resistance in the country.
Our Government is thus really a Government by Experts, but is
cleverly disguised in such a manner as to appear “broadbased upon
the People’s Will.” A similar camouflage might very properly be
adopted in the management of factories and industrial enterprises.
But the State is older than the Factory, and has learnt more wisdom.
The other primary function of government is the maintenance of
public order—in other words, the enforcement of the laws it has
enacted. Among the Medes and Persians, whose laws were never
altered, this was the main function of government. And in a non-
progressive world the task of keeping everyone “in his place” and
preventing him from encroaching on his neighbours might well fill the
whole sphere of government. The Anarchist ideal of the abolition of
all government is only possible if we can abolish the natural
tendency of all living things to expand and extend their sphere of
action; or if we can endow its neighbours with sufficient reserves of
energy to be able to oppose any undue expansion. If a motorist
knocks down a pedestrian, we discourage that undue extension of
his sphere by a fine or imprisonment. If we could endow every
pedestrian with the physical power of stopping a car, say, by raising
his hand; or if, alternatively, we could make him invulnerable,
indestructible, and untraversable, there would be no need for
prosecutions, and Anarchism would, in that particular case, become
a possible system.
The idea of a government being something superior to ordinary
humanity is somewhat ludicrous. In actual practice, government is
the servant of the public, and not its master. There are countless
cases of the process of law being used for private ends. The rule of
conduct among some powerful individuals and corporations is to “go
on until you are stopped,” in other words, to do what you like until
somebody objects, or until you are stopped by the law. The law is
thus used as a sort of indicator or “automatic cut-out” much as an
electrical engineer would use a safety-fuse. The main difference is
that the “blowing” of some particular fuses leads to explosions and
permanent damage, as when a crime is committed.
The business of government is hard and sometimes very exacting
work. One can imagine some misdemeanant of the future being
condemned to carry on the government, or some important function
of it, for so many months, as the most exacting form of hard labour.
This elementary fact has of late been recognized by most modern
parliaments in the payment of their members. Government should be
recognized as a profession and rewarded as such. The late Mr. W. T.
Stead’s alternative to Democracy was an “Autocracy tempered by
Assassination.” There is another alternative, viz., Bureaucracy
tempered by Emigration. It is the system practised in such
institutions as Proprietary clubs. Members are not worried to elect
committees and honorary officers. If they are satisfied with the
management they remain in the club. If they are not, they join some
other club. The same process on a larger scale led to the foundation
of the American colonies and the United States. It is largely at work
at the present day, but is complicated by all sorts of restrictions and
difficulties, the divergence of languages being one of the most
serious obstacles. As intercommunication increases, the natural
tendency to go where one can be happiest—ubi bene, ibi patria—will
no doubt come increasingly into action, and will be a wholesome
check upon the extravagances of cranky legislatures.
But I doubt whether there is any tendency at all of governments to
become less effective. Almost every advance of science and
invention makes the maintenance of public order and security easier.
The tracing of criminals by wireless telegraphy and broadcasting is a
striking illustration of the aid science can give to the police. Almost
everywhere science and invention are on the side of the established
order. Although every researcher and inventor is, in a sense, a
revolutionary, in that his work is likely to produce immense changes
in human activity, his general outlook tends towards aristocracy,
since he is imbued with the sense of the immense differences in the
personal equipment of individuals, which no equalitarian
sentimentality will ever wipe out.
The Farther Outlook.—So far, we have looked but little ahead, a
century at most. The prophet’s task becomes more arduous as the
time is extended. Historical guidance fails us. Familiar landmarks get
blurred and disappear. We are in danger of getting lost in a bog of
unreal speculation. Yet the task has often been essayed, and it is
necessary and desirable that it be essayed now and again. Let me
make my own humble attempt, in the light of what knowledge I have
acquired and what great thoughts I have encountered in many lands
and languages, and in discussions with many thinkers.
We must extend our time scale from centuries to millennia, and
from millennia to geological eras. Above all, we must take into
account not only the rapid advance of science and invention, but the
constant acceleration of that advance.
The consequence of that constant acceleration is that new
developments and achievements succeed each other with
bewildering rapidity. Hardly have we got accustomed to the idea of
telegraphy without wires when radio-telephony becomes an
accomplished fact, and within a few years there is a rich crop of
listeners with their wireless receiving sets counting by the million. An
entirely new form of publicity comes into being, and a speaker on
Savoy Hill is able to speak to an audience of millions and sway them
by his voice more effectively than he can do by cold print in the
newspapers.
And this is only a beginning. Communication will become closer
and more general. Already the earth is a network of lines and cables,
linking continent to continent. Soon a speaker will have the earth for
his sounding board and his hall of audience, and the privilege of
addressing the human race will be prized above a coronation.
Human sight and hearing will extend its range enormously, not only
in space, but in time also. For the cinema film and improvements in
the recording of sound will make it possible to make minute and
comprehensive records of past sights and sounds for future
reproduction, so that nothing of any value may be lost.
Other progress will go hand in hand with the rapid development of
“signalling” communications, such as telegraphs and the like. The
transport of goods and passengers will rapidly gain in speed,
comfort, and safety, until the whole earth becomes accessible to all.
It will not only become accessible, but also habitable. The tropics,
the original cradle of the human race, will once more be reclaimed
from our most formidable enemies of the insect-world and the ever-
present bacterium. The higher organism will assert its much-
contested supremacy over those minute organisms which owe their
influence to mere numerical superiority. Our descendants will pay an
afternoon’s visit to Timbuctoo or Mount Ararat, much as we should
visit the British Museum or the Lake District. Everybody will be a
globe-trotter, but the “globe” will not be confined to the ordinary
tourist resorts. It will include every part of the world, even the Poles.
And wherever they go they will find friendly voices, long familiar in
the home through the service of radio-telephony. There will, of
course, be an international auxiliary language, understood
everywhere, a language artificial in its structure—every literary
language is largely artificial—but utilizing those roots which have
already become part and parcel of all cultured languages. This will
not mean the displacement or loss of native languages which have
proved their title to survival by their literature.
War will not cease for perhaps a century or more. But it will finally
cease when the truth has sunk in that war is a loss to every
belligerent and to the whole world. Human rivalry and competition
will take other forms. There are many ways of killing men, women
and children, besides suffocating them with chlorine. If a tribe is to
be exterminated, nobody will be killed, but all its members will be
painlessly sterilized by X-rays or some such modern means, so that
the next generation will know them no more. It will be more humane
than the Biblical expedient of “dashing their children against a
stone.”
The mass of interconnections between human nations and
individuals will be like a closely-woven fabric. Even now, the digging
up of a city road reveals a tangled network of water pipes, electric
mains, gas pipes, and drain pipes suggestive of the dissection of an
animal body. It is but a faint foreshadowing of what is to come. The
substratum of life will become more and more complete as
conscious life becomes simpler.
Nobody is conscious of the appalling complexity of his anatomical
organization when using his body as a well-poised instrument of
thought and intercourse. “The simple life” is not the old-fashioned
country life of England or the primitive life of savage humanity. Real
simplicity is constituted by the life in which most things are done by
pressing a button, and a man can travel across a continent in such
comfort that on arrival at his destination all memories of his journey
are dimmed or lost, and he can hardly recall having travelled at all.
We may, therefore, expect that, as facilities for intercourse become
more detailed and widespread, the effect will be, not to increase the
tax on our nerve force until it becomes unbearable, but to increase
our area of selection. There will thus be more consistency in our
actual interests and activities and more real harmony and leisure.
The unification of the planet which is being accomplished before
our eyes will have some astounding consequences. Mankind will
assume a definite mastery of his home in the solar system. Attila
could boast that when he plunged his spear into the ground, the
whole earth trembled. The earth trembles even now to the electric
signals of our powerful wireless stations. What will it be in a hundred
or a thousand years? In a hundred years the unification of the
human race will be complete. The earth and the fulness thereof will
be under the full mastery of man. All animal, vegetable and bacterial
life will be kept within strict bounds in the interests of humanity. The
earth will be under one government, and one language will be written
and understood, or even spoken, all over the globe. There will still be
different races and perhaps allied nations, but travel and commerce
will be free and unfettered, and calamities will be alleviated and
dangers met by the united forces of all mankind.
And all the world will be young. The advances of medicine and
surgery will have been such that most of the ailments and limitations
of old age will have been eliminated. Life will be prolonged at its
maximum of efficiency until death comes like sunset, and is met
without pain and without reluctance. There will be no death from
disease, and almost any sort of injury will be curable.
And in a thousand years? What will become of our globe and its
dominant race, if no great catastrophe occurs to stop its exponential
curve of progress. But for that exponential curve and its tendency
towards constant acceleration, a thousand years would be no great
period to foretell. Life has become world-wide in the last thousand
years. The intellectual outlook has increased with the area of travel
and communication. Dogmas and shibboleths have lost their force.
Art and science have been emancipated from their ecclesiastical
fetters. But the immense leap made since coal came into its own as
a world-force belongs to our own age. The exhaustion of the coal-
fields might slow down progress for a time, but so long as mankind
keeps its continuity, its past achievements and its rate of
achievement will act as a stimulus and encouragement to further
efforts, and new sources of energy will be discovered and utilized.
And so we may feel justified in expecting continual progress for at
least a thousand years. Can we imagine the result? A globe laid out
like a huge garden, with a climate under perfect control; the internal
heat of the earth brought to the surface and utilized as a source of
never-failing energy. Portions of the interior of the earth reclaimed
and made habitable; all machinery and sources of power wisely
distributed and made instantly available for all legitimate purposes.
The earth’s surface and the rippling ether in which it swims made
into a vast playground of human thought and emotion, and all
mankind throbbing in unison to every great thought.
The Earth will have become a sentient being.—It will be as closely
unified and organized as the human individual himself. Mankind will
be the “grey matter” of its brain. It may not resemble a sentient being

Change Detection and Image Time Series Analysis 1 Unsupervised Methods 1St Edition Abdourrahmane M Atto Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Change Detection and Image Time Series Analysis 1 Unsupervised Methods 1St Edition Abdourrahmane M Atto Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Change Detection and Image Time

Series Analysis 1: Unsupervised

Methods and techniques for fire detection : signal,

Interrupted Time Series Analysis David Mcdowall

Applied economic forecasting using time series methods

Time Series Analysis in Seismology: Practical

Demanding Energy: Space, Time and Change 1st Edition

Machine vision inspection systems. Vol.1: image

Deep Learning for Medical Image Analysis 1st Edition S.

Applied Time Series Analysis. A Practical Guide to

Change Detection and Image

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2021

Library of Congress Control Number: 2021941648

British Library Cataloguing-in-Publication Data

Chapter 1. Unsupervised Change Detection in Multitemporal

Chapter 2. Change Detection in Time Series of Polarimetric

Chapter 3. An Overview of Covariance-based Change Detection

Chapter 4. Unsupervised Functional Information Clustering

Chapter 5. Thresholds and Distances to Better Detect Wet Snow

Chapter 6. Fractional Field Image Time Series Modeling

6.2.1. Cyclone texture feature . . . . . . . . . . . . . . . . . . . . . . . . . 149

Chapter 7. Graph of Characteristic Points for Texture Tracking:

Chapter 8. Multitemporal Analysis of Sentinel-1/2 Images

8.3. SAR processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

Chapter 9. Statistical Difference Models for Change Detection

List of Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275

Summary of Volume 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

Abdourrahmane M. ATTO1 , Francesca B OVOLO2

This book is part of the ISTE-Wiley “SCIENCES” Encyclopedia and belongs to

Change Detection and Image Time Series Analysis 1,

Volume 1: Unsupervised methods

A significant part of this book is dedicated to a wide range of unsupervised

Volume 2: Supervised methods

The second volume of this book is dedicated to supervised methods. Chapter 1 of

Change Detection and Image Time Series Analysis 1,

Sicong L IU1 , Francesca B OVOLO2 , Lorenzo B RUZZONE3 ,

Remote sensing satellites have a great potential to recurrently monitor the

Change Detection and Image Time Series Analysis 1,

is often very costly and labor-intensive. In particular, in some cases such as in

Basically, CD techniques are developed based on specific remote sensing satellite

In this chapter, we focus on the unsupervised CD problem in multitemporal

1.2. Unsupervised change detection in multispectral images

1.2.1. Related concepts

Depending on the purpose of unsupervised CD tasks, two main categories of

In general, the unsupervised CD process includes the following main steps:

Multitemporal Data Feature Generation Change Index Change Detection Performance

• Calibration • Spectral Features • Differencing • Change Number • Accuracy Index

Multitemporal data pre-processing: in this step, different operations such as

Feature generation and selection: features extracted from original multitemporal

Change detection algorithm design: unlike the supervised and semi-supervised

as empirical segmentation (Bruzzone and Prieto 2000b), Kittler–Illingworth (KI)

Performance evaluation: similar to the supervised CD methods, unsupervised

1.2.2. Open issues and challenges

The current development of unsupervised CD techniques for multispectral remote

1.2.3. Spectral–spatial unsupervised CD techniques

Despite the success of aforementioned CD methods, especially the CVA-based

To address the above problems in pixel-based CD (PBCD) techniques,

For OBCD methods, four main categories exist: image-object, class-object,

1.3. Unsupervised multiclass change detection approaches based on

1.3.1. Sequential spectral change vector analysis (S2 CVA)

Here, we first recall a popular state-of-the-art unsupervised multiclass CD method,

where E rXD s is the expectation of XD . The eigen-decomposition of A can be