Image Analysis, Classification, and Change Detection in Remote Sensing: With Algorithms for ENVI/IDL, Second Edition

Demonstrating the breadth and depth of growth in the field since the publication of the popular first edition, Image Analysis, Classification and Change Detection in Remote Sensing, with Algorithms for ENVI/IDL, Second Edition has been updated and expanded to keep pace with the latest versions of the ENVI software environment. Effectively interweaving theory, algorithms, and computer codes, the text supplies an accessible introduction to the techniques used in the processing of remotely sensed imagery.

This significantly expanded edition presents numerous image analysis examples and algorithms, all illustrated in the array-oriented language IDL―allowing readers to plug the illustrations and applications covered in the text directly into the ENVI system―in a completely transparent fashion. Revised chapters on image arrays, linear algebra, and statistics convey the required foundation, while updated chapters detail kernel methods for principal component analysis, kernel-based clustering, and classification with support vector machines.

Additions to this edition include:

• An introduction to mutual information and entropy
• Algorithms and code for image segmentation
• In-depth treatment of ensemble classification (adaptive boosting )
• Improved IDL code for all ENVI extensions, with routines that can take advantage of the parallel computational power of modern graphics processors
• Code that runs on all versions of the ENVI/IDL software environment from ENVI 4.1 up to the present―available on the author's website
• Many new end-of-chapter exercises and programming projects

With its numerous programming examples in IDL and many applications supporting ENVI, such as data fusion, statistical change detection, clustering and supervised classification with neural networks―all available as downloadable source code―this self-contained text is ideal for classroom use or self study.

Images, Arrays, and Matrices
Multispectral Satellite Images
Algebra of Vectors and Matrices
Eigenvalues and Eigenvectors
Singular Value Decomposition
Vector Derivatives
Finding Minima and Maxima

Image Statistics
Random Variables
Random Vectors
Parameter Estimation
Hypothesis Testing and Sample Distribution Functions
Conditional Probabilities, Bayes’ Theorem, and Classification
Ordinary Linear Regression
Entropy and Information

Transformations
Discrete Fourier Transform
Discrete Wavelet Transform
Principal Components
Minimum Noise Fraction
Spatial Correlation

Filters, Kernels, and Fields
Convolution Theorem
Linear Filters
Wavelets and Filter Banks
Kernel Methods
Gibbs–Markov Random Fields

Image Enhancement and Correction
Lookup Tables and Histogram Functions
Filtering and Feature Extraction
Panchromatic Sharpening
Topographic Correction
Image–Image Registration

Supervised Classification: Part 1
Maximum a Posteriori Probability
Training Data and Separability
Maximum Likelihood Classification
Gaussian Kernel Classification
Neural Networks
Support Vector Machines

Supervised Classification: Part 2
Postprocessing
Evaluation and Comparison of Classification Accuracy
Adaptive Boosting
Hyperspectral Analysis

Unsupervised Classification
Simple Cost Functions
Algorithms That Minimize the Simple Cost Functions
Gaussian Mixture Clustering
Including Spatial Information
Benchmark
Kohonen Self-Organizing Map
Image Segmentation

Change Detection
Algebraic Methods
Postclassification Comparison
Principal Components Analysis
Multivariate Alteration Detection
Decision Thresholds and Unsupervised Classification of Changes
Radiometric Normalization

Appendix A: Mathematical Tools
Cholesky Decomposition
Vector and Inner Product Spaces
Least Squares Procedures

Appendix B: Efficient Neural Network Training Algorithms
Hessian Matrix
Scaled Conjugate Gradient Training
Kalman Filter Training
A Neural Network Classifier with Hybrid Training

Appendix C: ENVI Extensions in IDL
Installation
Extensions

Appendix D: Mathematical Notation

References

Index