Muutke küpsiste eelistusi

E-raamat: Kernel Methods For Remote Sensing Data Analysis [Wiley Online]

Edited by (University of Valencia, Spain), Edited by (University of Trento, Italy)
  • Formaat: 434 pages
  • Ilmumisaeg: 23-Oct-2009
  • Kirjastus: John Wiley & Sons Inc
  • ISBN-10: 470748990
  • ISBN-13: 9780470748992
Teised raamatud teemal:
  • Wiley Online
  • Hind: 163,88 €*
  • * hind, mis tagab piiramatu üheaegsete kasutajate arvuga ligipääsu piiramatuks ajaks
Kernel Methods For Remote Sensing Data AnalysisSuurem pilt
  • Formaat: 434 pages
  • Ilmumisaeg: 23-Oct-2009
  • Kirjastus: John Wiley & Sons Inc
  • ISBN-10: 470748990
  • ISBN-13: 9780470748992
Teised raamatud teemal:
Wiley Online e-raamatudRohkem infot Wiley Online kohta
Raamatu kodulehekülg: https://onlinelibrary.wiley.com/doi/book/10.1002/9780470748992
This work presents research related to remote sensing techniques based on recent advances in kernel methods. Two early chapters provide background on machine learning techniques in remote sensing data analysis, and theoretical and practical foundations of kernel methods. The rest of book addresses recent research in developing kernel methods in remote sensing for supervised classification, semi-supervised classification, regression, and feature extraction. Some specific topics covered include the support vector machine (SVM) algorithm, a domain adaptation SVM for land-cover map updating, kernel methods for unmixing hyperspectral imagery, kernel-based quantitative remote sensing inversion, and kernel multivariate analysis in remote sensing feature extraction. The book is for engineers, scientists, and researchers involved in remote sensing data processing. Camps-Valls teaches in the Department of Electronics Engineering at the University of Valencia, Spain. Bruzzone is professor of telecommunications at the University of Trento, Italy. Annotation ©2010 Book News, Inc., Portland, OR (booknews.com)

Kernel methods have long been established as effective techniques in the framework of machine learning and pattern recognition, and have now become the standard approach to many remote sensing applications. With algorithms that combine statistics and geometry, kernel methods have proven successful  across many different domains related to the analysis of images of the Earth acquired from airborne and satellite sensors, including natural resource control, detection and monitoring of anthropic infrastructures (e.g. urban areas), agriculture inventorying, disaster prevention and damage assessment, and anomaly and target detection.

 

Presenting the theoretical foundations of kernel methods (KMs) relevant to the remote sensing domain, this book serves as a practical guide to the design and implementation of these methods. Five distinct parts present state-of-the-art research related to remote sensing based on the recent advances in kernel methods, analysing the related methodological and practical challenges:

  • Part I introduces the key concepts of machine learning for remote sensing, and the theoretical and practical foundations of kernel methods.
  • Part II explores supervised image classification including Super Vector Machines (SVMs), kernel discriminant analysis, multi-temporal image classification, target detection with kernels, and Support Vector Data Description (SVDD) algorithms for anomaly detection.
  • Part III looks at semi-supervised classification with transductive SVM approaches for hyperspectral image classification and kernel mean data classification.
  • Part IV examines regression and model inversion, including the concept of a kernel unmixing algorithm for hyperspectral imagery, the theory and methods for quantitative remote sensing inverse problems with kernel-based equations, kernel-based BRDF (Bidirectional Reflectance Distribution Function), and temperature retrieval KMs. 
  • Part V deals with kernel-based feature extraction and provides a review of the principles of several multivariate analysis methods and their kernel extensions.

This book is aimed at engineers, scientists and researchers involved in remote sensing data processing, and also those working within machine learning and pattern recognition.

About the editors xiii
List of authors
xv
Preface xix
Acknowledgments xxiii
List of symbols
xxv
List of abbreviations
xxvii
I Introduction
1(48)
Machine learning techniques in remote sensing data analysis
3(22)
Bjorn Waske
Mathieu Fauvel
Jon Atli Benediktsson
Jocelyn Chanussot
Introduction
3(7)
Challenges in remote sensing
3(1)
General concepts of machine learning
4(2)
Paradigms in remote sensing
6(4)
Supervised classification: algorithms and applications
10(10)
Bayesian classification strategy
10(1)
Neural networks
11(2)
Support Vector Machines (SVM)
13(4)
Use of multiple classifiers
17(3)
Conclusion
20(1)
Acknowledgments
21(1)
References
21(4)
An introduction to kernel learning algorithms
25(24)
Peter V. Gehler
Bernhard Scholkopf
Introduction
25(1)
Kernels
26(10)
Measuring similarity with kernels
26(1)
Positive definite kernels
27(2)
Constructing the reproducing kernel Hilbert space
29(2)
Operations in RKHS
31(1)
Kernel construction
32(1)
Examples of kernels
33(3)
The representer theorem
36(1)
Learning with kernels
37(8)
Support vector classification
38(1)
Support vector regression
39(1)
Gaussian processes
39(1)
Multiple kernel learning
40(2)
Structured prediction using kernels
42(1)
Kernel principal component analysis
43(1)
Applications of support vector algorithms
44(1)
Available software
44(1)
Conclusion
45(1)
References
45(4)
II Supervised image classification
49(144)
The Support Vector Machine (SVM) algorithm for supervised classification of hyperspectral remote sensing data
51(34)
J. Anthony Gualtieri
Introduction
52(1)
Aspects of hyperspectral data and its acquisition
53(3)
Hyperspectral remote sensing and supervised classification
56(1)
Mathematical foundations of supervised classification
57(6)
Empirical risk minimization
58(1)
General bounds for a new risk minimization principle
58(3)
Structural risk minimization
61(2)
From structural risk minimization to a support vector machine algorithm
63(7)
SRM for hyperplane binary classifiers
63(1)
SVM algorithm
64(2)
Kernel method
66(2)
Hyperparameters
68(1)
A toy example
68(1)
Multi-class classifiers
68(1)
Data centring
69(1)
Benchmark hyperspectral data sets
70(2)
The 4 class subset scene
70(1)
The 16 class scene
71(1)
The 9 class scene
71(1)
Results
72(5)
SVM implementation
72(1)
Effect of hyperparameter d
72(1)
Measure of accuracy of results
73(1)
Classifier results for the 4 class subset scene and the 16 class full scene
74(1)
Results for the 9 class scene and comparison of SVM with other classifiers
74(1)
Effect of training set size
75(1)
Effect of simulated noisy data
75(2)
Using spatial coherence
77(1)
Why do SVMs perform better than other methods?
78(1)
Conclusions
79(1)
References
79(6)
On training and evaluation of SVM for remote sensing applications
85(26)
Giles M. Foody
Introduction
85(1)
Classification for thematic mapping
86(2)
Overview of classification by a SVM
88(2)
Training stage
90(7)
General recommendations on sample size
91(3)
Training a SVM
94(3)
Summary on training
97(1)
Testing stage
97(6)
General issues in testing
98(5)
Specific issues for SVM classification
103(1)
Conclusion
103(1)
Acknowledgments
104(1)
References
104(7)
Kernel Fisher's Discriminant with heterogeneous kernels
111(14)
M. Murat Dundar
Glenn Fung
Introduction
111(1)
Linear Fisher's Discriminant
112(2)
Kernel Fisher Discriminant
114(2)
Mathematical programming formulation
114(2)
Kernel Fisher's Discriminant with heterogeneous kernels
116(2)
Automatic kernel selection KFD algorithm
118(1)
Numerical results
119(4)
Dataset used: Purdue Campus data
119(1)
Classifier design
120(1)
Analysis of the results
121(2)
Conclusion
123(1)
References
123(2)
Multi-temporal image classification with kernels
125(22)
Jordi Munoz-Mari
Luis Gomez-Chova
Manel Martinez-Ramon
Jose Luis Rojo-Alvarez
Javier Calpe-Maravilla
Gustavo Camps-Valls
Introduction
126(3)
Multi-temporal classification methods
126(1)
Change detection methods
127(1)
The proposed kernel-based framework
128(1)
Multi-temporal classification and change detection with kernels
129(5)
Problem statement and notation
129(1)
Mercer's kernels properties
130(1)
Composite kernels for multi-temporal classification
131(2)
Composite kernels for change detection
133(1)
Contextual and multi-source data fusion with kernels
134(1)
Composite kernels for integrating contextual information
134(1)
Composite kernels for dealing with multi-source data
134(1)
Remarks
134(1)
Multi-temporal/-source urban monitoring
135(6)
Model development and free parameter selection
135(1)
Data collection and feature extraction
135(3)
Multi-temporal image classification
138(1)
Change detection
138(3)
Classification maps
141(1)
Conclusions
141(2)
Acknowledgments
143(1)
References
143(4)
Target detection with kernels
147(22)
Nasser M. Nasrabadi
Introduction
147(2)
Kernel learning theory
149(1)
Linear subspace-based anomaly detectors and their kernel versions
150(11)
Principal component analysis
151(1)
Kernel PCA subspace-based anomaly detection
152(2)
Fisher linear discriminant analysis
154(1)
Kernel fisher discriminant analysis
154(2)
Eigenspace separation transform
156(1)
Kernel eigenspace separation transform
157(2)
RX algorithm
159(1)
Kernel RX algorithm
160(1)
Results
161(5)
Simulated toy data
162(1)
Hyperspectral imagery
163(3)
Conclusion
166(1)
References
166(3)
One-class SVMs for hyperspectral anomaly detection
169(24)
Amit Banerjee
Philippe Burlina
Chris Diehl
Introduction
169(3)
Deriving the SVDD
172(4)
The linear SVDD
172(1)
The kernel-based SVDD
173(3)
SVDD function optimization
176(1)
SVDD algorithms for hyperspectral anomaly detection
177(6)
Outline of algorithms
177(2)
Dimensions for the background window
179(1)
Estimating sigma
179(2)
Normalized SVDD test statistic
181(2)
Experimental results
183(7)
Conclusions
190(1)
References
191(2)
III Semi-supervised image classification
193(54)
A domain adaptation SVM and a circular validation strategy for land-cover maps updating
195(28)
Mattia Marconcini
Lorenzo Bruzzone
Introduction
195(3)
Literature survey
198(2)
Learning under sample selection bias: transductive and semi-supervised methods
198(2)
Domain adaptation: partially-unsupervised methods
200(1)
Proposed domain adaptation SVM
200(8)
DASVM: problem definition and assumptions
201(1)
DASVM: formulation
201(7)
Proposed circular validation strategy
208(2)
Circular validation strategy: rationale
208(1)
Circular validation strategy: formulation
209(1)
Experimental results
210(8)
Discussions and conclusion
218(1)
References
219(4)
Mean kernels for semi-supervised remote sensing image classification
223(24)
Luis Gomez-Chova
Javier Calpe-Maravilla
Lorenzo Bruzzone
Gustavo Camps-Valls
Introduction
224(1)
Semi-supervised classification with mean kernels
225(7)
Learning from labelled samples
225(1)
Image clustering
226(1)
Cluster similarity and the mean map
226(2)
Composite sample-cluster kernels
228(1)
Sample selection bias and the soft mean map
229(2)
Summary of composite mean kernel methods
231(1)
Experimental results
232(11)
Model development
232(1)
Results on synthetic data
232(1)
Results on real data
233(10)
Conclusions
243(1)
Acknowledgments
243(1)
References
244(3)
IV Function approximation and regression
247(80)
Kernel methods for unmixing hyperspectral imagery
249(22)
Joshua Broadwater
Amit Banerjee
Philippe Burlina
Introduction
249(1)
Mixing models
250(2)
Areal mixtures
251(1)
Intimate mixtures
251(1)
Proposed kernel unmixing algorithm
252(6)
Support vector data description for endmember extraction
254(1)
Rate-distortion theory
255(1)
Kernel fully constrained least squares abundance estimates
256(2)
Outline of full algorithm
258(1)
Experimental results of the kernel unmixing algorithm
258(7)
RELAB data results
259(2)
AVIRIS data results
261(3)
Processing times
264(1)
Development of physics-based kernels for unmixing
265(1)
Simplification of the albedo to reflectance transform
265(1)
Kernel approximation of intimate mixtures
265(1)
Physics-based kernel results
266(2)
Summary
268(1)
References
268(3)
Kernel-based quantitative remote sensing inversion
271(30)
Yanfei Wang
Changchun Yang
Xiaowen Li
Introduction
272(1)
Typical kernel-based remote sensing inverse problems
273(3)
Aerosol inverse problems
274(1)
Land surface parameter retrieval problem
275(1)
Well-posedness and ill-posedness
276(2)
Regularization
278(7)
Imposing a priori constraints on the solution
278(1)
Tikhonov variational regularization
278(4)
Direct regularization
282(2)
Statistical regularization
284(1)
Optimization techniques
285(3)
Sparse inversion in l1 space
285(1)
Optimization methods for l2 minimization model
286(2)
Kernel-based BRDF model inversion
288(5)
Inversion by NTSVD
288(1)
Tikhonov regularized solution
288(1)
Land surface parameter retrieval results
289(4)
Aerosol particle size distribution function retrieval
293(3)
Conclusion
296(1)
Acknowledgments
296(1)
References
296(5)
Land and sea surface temperature estimation by support vector regression
301(26)
Gabriele Moser
Sebastiano B. Serpico
Introduction
302(1)
Previous work
303(3)
LST and SST estimation from satellite data
303(2)
Parameter optimization and error modelling for SVR
305(1)
Methodology
306(5)
SVR for LST and SST estimation
306(1)
Automatic parameter optimization for SVR
307(2)
Pointwise statistical modelling the SVR error
309(2)
Experimental results
311(9)
Data sets and experimental set-up
311(2)
Parameter-optimization results
313(5)
Results on the estimation of regression-error variance
318(2)
Conclusions
320(2)
Acknowledgments
322(1)
References
322(5)
V Kernel-based feature extraction
327(74)
Kernel multivariate analysis in remote sensing feature extraction
329(24)
Jeronimo Arenas-Garcia
Kaare Brandt Petersen
Introduction
329(3)
Multivariate analysis methods
332(7)
Principal component analysis (PCA)
333(2)
Partial least squares
335(2)
Canonical correlation analysis
337(1)
Orthonormalized partial least squares
338(1)
Kernel multivariate analysis
339(5)
Kernel PCA
340(1)
Kernel PLS
341(1)
Kernel CCA
342(1)
Kernel OPLS
343(1)
Some considerations about Kernel MVA methods
344(1)
Sparse Kernel OPLS
344(2)
Experiments: pixel-based hyperspectral image classification
346(4)
Data set description and experimental setup
346(1)
Results description
347(3)
Conclusions
350(1)
Acknowledgments
351(1)
References
351(2)
KPCA algorithm for hyperspectral target/anomaly detection
353(22)
Yanfeng Gu
Introduction
353(1)
Motivation
354(3)
Feature extraction of hyperspectral images
354(1)
Introducing KM for hyperspectral image processing
355(1)
Hyperspectral images for numerical experiments
356(1)
Kernel-based feature extraction in hyperspectral images
357(3)
Principal component analysis
357(1)
Kernel mapping
358(1)
Kernel Principal Component Analysis (KPCA)
358(2)
Kernel-based target detection in hyperspectral images
360(4)
The concept of target detection
361(1)
Invariant subpixel material detector
361(1)
Kernel invariant subpixel detection
362(2)
Kernel-based anomaly detection in hyperspectral images
364(8)
The concept of anomaly detection
364(2)
RX detector
366(1)
Selective KPCA Feature Extraction for Anomaly Detection
367(5)
Conclusions
372(1)
Acknowledgments
372(1)
References
372(3)
Remote sensing data classification with kernel nonparametric feature extractions
375(26)
Bor-Chen Kuo
Jinn-Min Yang
Cheng-Hsuan Li
Introduction
376(1)
Related feature extractions
377(6)
Linear discriminant analysis
377(1)
Generalized discriminant analysis
378(2)
Nonparametric weighted feature extraction
380(2)
Fuzzy linear feature extraction
382(1)
Kernel-based NWFE and FLFE
383(5)
Kernel-based NWFE
383(3)
Kernel-based FLFE
386(2)
Eigenvalue resolution with regularization
388(1)
Experiments
389(9)
Data sets
389(3)
Experiment design
392(1)
Experiment results
392(6)
Comments and conclusions
398(1)
References
398(3)
Index 401
Gustavo Camps-Valls was born in Valencia, Spain in 1972, and received a B.Sc. degree in Physics (1996), a B.Sc. degree in Electronics Engineering (1998), and a Ph.D. degree in Physics (2002) from the Universitat de Valencia. He is currently an associate professor in the Department of Electronics Engineering at the Universitat de Valencia, where he teaches electronics, advanced time series processing, machine learning for remote sensing and digital signal processing. His research interests are tied to the development of machine learning algorithms for signal and image processing, with special attention to adaptive systems, neural networks and kernel methods. He conducts and supervises research on the application of these methods to remote sensing image analysis and recognition, and image denoising and coding. Dr Camps-Valls is the author (or co-author) of 50 papers in referred international journals, more than 70 international conference papers, 15 book chapters, and is editor of other related books, such as Kernel Methods in Bioengineering, Signal and Image Processing (IGI, 2007). He has served as reviewer to many international journals, and on the Program Committees of SPIE Europe, IGARSS, IWANN and ICIP. Dr Camps-Valls was a member of the European Network on Intelligent Technologies for Smart Adaptive Systems (EUNITE), and the Spanish Thematic Networks on 'Pattern Recognition' and 'Biomedical Engineering'. He is active in the R+D sector through a large number of projects funded by both public and industrial partners, both at national and international levels. He is an Evaluator of project proposals and scientific organizations. Since 2003 he has been a member of the IEEE and SPIE. Since 2009 he has been a member of the machine Learning for Signal Processing (MLSP) Technical Committee of the IEEE Signal Processing Society. Visit http://www.uv.es/gcamps for more information. Lorenzo Bruzzone received a laurea (M.S.) degree in electronic engineering (summa cum laude) ad a Ph.D. degree in telecommunications from the University of Genoa, Italy, in 1993 and 1998, respectively. From 1998 to 2000 he was a Postdoctoral researcher at the University of Genoa. In 2000 he joined the University of Trento, Italy, where he is currently a Full Professor telecommunications. He teaches remote sensing, pattern recognition, radar and electrical communications. Dr Bruzzone is the Head of the remote Sensing Laboratory in the Department of Information Engineering and Computer Science, University of Trento. His current research interests are in the area of remote-sensing image processing and recognition (analysis of multitemporal data, feature extraction and election, classification, regression and estimation, data fusion and machine learning). He conducts and supervises research on these topics within the frameworks of several national and international projects. He is an Evaluator of project proposals for many different governments (including the European Commission) and scientific organizations. He is the author (or co-author) of 74 scientific publication in referred international journals, more than 140 papers in conference proceedings and 7 book chapters. He is a referee for many international journals and has served on the Scientific Committees of several international conferences. He is a member of the Managing Committee of the Italian Inter-University Consortium on Telecommunications and a member of the Scientific Committee of the India-Italy Center for Advanced Research. Since 2009 he has been a member of the Administrative Committee of the IEEE Geoscience and Remote Sensing Society. Dr Bruzzone gained first place in the Student Prize Paper Competition of the 1998 IEEE International Geoscience and Remote Sensing Symposium (Seattle, July 1998). He was a recipient of the Recognition of IEEE Transactions on Geoscience and remote Sensing Best reviewers in 1999 and was a Guest Editor of a Special Issue of the IEEE Transactions on Geoscience and Remote Sensing on the subject of the analysis of multitemporal remote-sensing images (November 2003). He was the General Chair and Co-chair of the First and Second IEEE International Workshop on the Analysis of Multi-temporal remote-Sensing Images (MultiTemp), and is currently a member of the Permanent Steering Committee of this series of workshops. Since 2003, he has been the Chair of the SPIE Conference on Image and Signal Processing for Remote Sensing. From 2004 to 2006 he served as an Associate Editor for the IEEE Geoscience and Remote Sensing Letters, and currently is an Associate Editor for the IEEE Transactions on Geoscience and Remote Sensing, and the Canadian Journal of Remote Sensing. He is a Senior member of IEEE, and also a member of the International Association for Pattern Recognition and of the Italian Association for Remote Sensing (AIT).
Püsilink: https://www.kriso.ee/db/9780470748992_pe.html
Märksõnad: