Remote Sensing Digital Image Analysis: An Introduction

by Richards, John A.

• ISBN13: 9783540648604
• ISBN10: 3540648607
• Edition: 3rd
• Format: Hardcover
• Copyright: 1999-06-01
• Publisher: SPRINGER

Summary

Remote Sensing Digital Image Analysis provides the non-specialist with an introduction to quantitative evaluation of satellite and aircraft derived remotely retrieved data. Each chapter covers the pros and cons of digital remotely sensed data, without detailed mathematical treatment of computer based algorithms, but in a manner conductive to an understanding of their capabilities and limitations. Problems conclude each chapter. This fourth edition has been developed to reflect the changes that have occurred in this area over the past several years. Its focus is on those procedures that seem now to have become part of the set of tools regularly used to perform thematic mapping. As with previous revisions, the fundamental material has been preserved in its original form because of its tutorial value; its style has been revised in places and it has been supplemented if newer aspects have emerged in the time since the third edition appeared. It still meets, however, the needs of the senior student and practitioner.

Table of Contents

Sources and Characteristics of Remote Sensing Image Data	1	(38)
Introduction to Data Sources	1	(6)
Characteristics of Digital Image Data	1	(1)
Spectral Ranges Commonly Used in Remote Sensing	2	(4)
Concluding Remarks	6	(1)
Weather Satellite Sensors	7	(2)
Polar Orbiting and Geosynchronous Satellites	7	(1)
The NOAA AVHRR (Advanced Very High Resolution Radiometer)	8	(1)
The Nimbus CZCS (Coastal Zone Colour Scanner)	8	(1)
GMS VISSR (Visible and Infrared Spin Scan Radiometer) and GOES Imager	8	(1)
Earth Resource Satellite Sensors in the Visible and Infrared Regions	9	(10)
The Landsat System	9	(1)
The Landsat Instrument Complement	10	(1)
The Return Beam Vidicon (RBV)	11	(1)
The Multispectral Scanner (MSS)	11	(2)
The Thematic Mapper (TM) and Enhanced Thematic Mapper+ (ETM+)	13	(1)
The SPOT HRV, HRVIR, HRG and Vegetation Instruments	14	(1)
ADEOS (Advanded Earth Observing Satellite)	15	(1)
Sea-Viewing Wide Field of View Sensor (SeaWiFS)	16	(1)
Marine Observation Satellite (MOS)	16	(2)
Indian Remote Sensing Satellite (IRS)	18	(1)
RESURS-O1	18	(1)
Aircraft Scanners in the Visible and Infrared Regions	19	(6)
General Considerations	19	(1)
The Daedalus AADS 1240/1260 Multispectral Line Scanner	20	(1)
The Airborne Thematic Mapper (ATM)	21	(1)
The Thermal Infrared Multispectral Scanner (TIMS)	21	(1)
Imaging Spectrometers	22	(3)
Image Data Sources in the Microwave Region	25	(5)
Side Looking Airborne Radar and Synthetic Aperture Radar	25	(2)
The Seasat SAR	27	(1)
Spaceborne (Shuttle) Imaging Radar-A (SIR-A)	28	(1)
Spaceborne (Shuttle) Imaging Radar-B (SIR-B)	28	(1)
Spaceborne (Shuttle) Imaging Radar-C (SIR-C)/X-band Synthetic Aperture Radar (X-SAR)	28	(1)
ERS-1, 2	28	(1)
JERS-1	28	(1)
Radarsat	29	(1)
Aircraft Imaging Radar Systems	29	(1)
Spatial Data Sources in General	30	(5)
Types of Spatial Data	30	(1)
Data Formats	31	(1)
Geographic Information Systems (GIS)	32	(2)
The Challenge to Image Processing and Analysis	34	(1)
A Comparison of Scales in Digital Image Data	35	(4)
References for Chapter 1	36	(1)
Problems	37	(2)
Error Correction and Registration of Image Data	39	(36)
Sources of Radiometric Distortion	39	(5)
The Effect of the Atmosphere on Radiation	39	(4)
Atmospheric Effects on Remote Sensing Imagery	43	(1)
Instrumentation Errors	43	(1)
Correction of Radiometric Distortion	44	(4)
Detailed Correction of Atmospheric Effects	44	(2)
Bulk Correction of Atmospheric Effects	46	(1)
Correction of Instrumentation Errors	47	(1)
Sources of Geometric Distortion	48	(8)
Earth Rotation Effects	49	(2)
Panoramic Distortion	51	(2)
Earth Curvature	53	(1)
Scan Time Skew	54	(1)
Variations in Platform Altitude, Velocity and Attitude	54	(1)
Aspect Radio Distortion	55	(1)
Sensor Scan Nonlinearities	55	(1)
Correction of Geometric Distortion	56	(10)
Use of Mapping Polynomials for Image Correction	56	(1)
Mapping Polynomials and Ground Control Points	57	(1)
Resampling	58	(1)
Interpolation	58	(2)
Choice of Control Points	60	(1)
Example of Registration to a Map Grid	61	(3)
Mathematical Modelling	64	(1)
Aspect Ratio Correction	64	(1)
Earth Rotation Skew Correction	64	(1)
Image Orientation to North-South	65	(1)
Correction of Panoramic Effects	65	(1)
Combining the Corrections	65	(1)
Image Registration	66	(4)
Georeferencing and Geocoding	66	(1)
Image to Image Registration	66	(1)
Sequential Similarity Detection Algorithm	67	(1)
Example of Image to Image Registration	67	(3)
Miscellaneous Image Geometry Operations	70	(5)
Image Rotation	70	(1)
Scale Changing and Zooming	70	(1)
References for Chapter 2	71	(1)
Problems	71	(4)
The Interpretation of Digital Image Data	75	(14)
Approaches to Interpretation	75	(1)
Forms of Imagery for Photointerpretation	76	(3)
Computer Processing for Photointerpretation	79	(2)
An Introduction to Quantitative Analysis -- Classification	81	(1)
Multispectral Space and Spectral Classes	82	(2)
Quantitative Analysis by Pattern Recognition	84	(5)
Pixel Vectors and Labelling	84	(1)
Unsupervised Classification	85	(1)
Supervised Classification	85	(2)
References for Chapter 3	87	(1)
Problems	88	(1)
Radiometric Enhancement Techniques	89	(24)
Introduction	89	(2)
Point Operations and Look Up Tables	89	(1)
Scalar and Vector Images	89	(2)
The Image Histogram	91	(1)
Contrast Modification in Image Data	92	(5)
Histogram Modification Rule	92	(1)
Linear Contrast Enhancement	93	(2)
Saturating Linear Contrast Enhancement	95	(1)
Automatic Contrast Enhancement	96	(1)
Logarithmic and Exponential Contrast Enhancement	96	(1)
Piecewise Linear Contrast Modification	96	(1)
Histogram Equalization	97	(5)
Use of the Cumulative Histogram	97	(4)
Anomalies in Histogram Equalization	101	(1)
Histogram Matching	102	(4)
Principle of Histogram Matching	102	(1)
Image to Image Contrast Matching	103	(1)
Matching to a Mathematical Reference	104	(2)
Density Slicing	106	(7)
Black and White Density Slicing	106	(1)
Colour Density Slicing and Pseudocolouring	107	(2)
References for Chapter 4	109	(1)
Problems	110	(3)
Geometric Enhancement Using Image Domain Techniques	113	(20)
Neighbourhood Operations	113	(1)
Template Operators	113	(1)
Geometric Enhancement as a Convolution Operation	114	(3)
Image Domain Versus Fourier Transformation Approaches	117	(1)
Image Smoothing (Low Pass Filtering)	118	(4)
Mean Value Smoothing	118	(3)
Median Filtering	121	(1)
Edge Detection and Enhancement	122	(7)
Linear Edge Detecting Templates	123	(1)
Spatial Derivative Techniques	124	(1)
The Roberts Operator	124	(1)
The Sobel Operator	125	(1)
Thinning, Linking and Border Responses	125	(1)
Edge Enhancement by Subtractive Smoothing (Sharpening)	126	(3)
Line Detection	129	(1)
Linear Line Detecting Templates	129	(1)
Non-linear and Semi-linear Line Detecting Templates	129	(1)
General Convolution Filtering	130	(1)
Shape Detection	130	(3)
References for Chapter 5	131	(1)
Problems	132	(1)
Multispectral Transformations of Image Data	133	(22)
The Principal Components Transformation	133	(15)
The Mean Vector and Covariance Matrix	133	(3)
A Zero Correlation, Rotational Transform	136	(6)
An Example -- Some Practical Considerations	142	(2)
The Effect of an Origin Shift	144	(1)
Application of Principal Components in Image Enhancement and Display	144	(1)
The Taylor Method of Contrast Enhancement	145	(3)
Other Applications of Principal Components Analysis	148	(1)
The Kauth-Thomas Tasseled Cap Transformation	148	(4)
Image Arithmetic, Band Ratios and Vegetation Indices	152	(3)
References of Chapter 6	152	(1)
Problems	153	(2)
Fourier Transformation of Image Data	155	(26)
Introduction	155	(1)
Special Functions	155	(3)
The Complex Exponential Function	155	(1)
The Dirac Delta Function	156	(1)
Properties of the Delta Function	157	(1)
The Heaviside Step Function	157	(1)
Fourier Series	158	(1)
The Fourier Transform	159	(1)
Convolution	160	(2)
The Convolution Integral	160	(2)
Convolution with an Impulse	162	(1)
The Convolution Theorem	162	(1)
Sampling Theory	162	(3)
The Discrete Fourier Transform	165	(8)
The Discrete Spectrum	165	(1)
Discrete Fourier Transform Formulae	166	(1)
Properties of the Discrete Fourier Transform	167	(1)
Computation of the Discrete Fourier Transform	168	(1)
Development of the Fast Fourier Transform Algorithm	168	(3)
Computational Cost of the Fast Fourier Transform	171	(1)
Bit Shuffling and Storage Considerations	172	(1)
The Discrete Fourier Transform of an Image	173	(4)
Definition	173	(1)
Evaluation of the Two Dimensional, Discrete Fourier Transform	173	(1)
The Concept of Spatial Frequency	174	(2)
Image Filtering for Geometric Enhancement	176	(1)
Convolution in Two Dimensions	176	(1)
Concluding Remarks	177	(4)
References for Chapter 7	178	(1)
Problems	179	(2)
Supervised Classification Techniques	181	(42)
Steps in Supervised Classification	181	(1)
Maximum Likelihood Classification	182	(7)
Bayes' Classification	182	(1)
The Maximum Likelihood Decision Rule	183	(1)
Multivariate Normal Class Models	184	(1)
Decision Surfaces	184	(1)
Thresholds	185	(2)
Number of Training Pixels Required for Each Class	187	(1)
A Simple Illustration	187	(2)
Minimum Distance Classification	189	(3)
The Case of Limited Training Data	189	(1)
The Discriminant Function	189	(1)
Degeneration of Maximum Likelihood to Minimum Distance Classification	190	(1)
Decision Surfaces	191	(1)
Thresholds	192	(1)
Parallelepiped Classification	192	(1)
Classification Time Comparison of the Classifiers	193	(1)
The Mahalanobis Classifier	194	(1)
Table Look Up Classification	194	(1)
Context Classification	195	(6)
The Concept of Spatial Context	195	(1)
Context Classification by Image Pre-Processing	195	(1)
Post Classification Filtering	196	(1)
Probabilistic Label Relaxation	196	(1)
The Basic Algorithm	196	(2)
The Neighbourhood Function	198	(1)
Determining the Compatibility Coefficients	199	(1)
The Final Step -- Stopping the Process	199	(1)
Examples	200	(1)
Classification Using Neural Networks	201	(22)
Linear Discrimination	201	(1)
Concept of a Weight Vector	201	(2)
Testing Class Membership	203	(1)
Training	204	(2)
Setting the Correction Increment	206	(1)
Classification -- The Threshold Logic Unit	206	(1)
Multicategory Classification	207	(1)
Networks of Classifiers -- Solutions of Nonlinear Problems	208	(1)
The Neural Networks Approach	209	(1)
The Processing Element	209	(2)
Training the Neural Network -- Backpropagation	211	(3)
Choosing the Network Parameters	214	(1)
Examples	215	(4)
References for Chapter 8	219	(2)
Problems	221	(2)
Clustering and Unsupervised Classification	223	(16)
Delineation of Spectral Classes	223	(1)
Similarity Metrics and Clustering Criteria	223	(2)
The Iterative Optimization (Migrating Means) Clustering Algorithm	225	(3)
The Basic Algorithm	225	(2)
Mergings and Deletions	227	(1)
Splitting Elongated Clusters	227	(1)
Choice of Initial Cluster Centres	227	(1)
Clustering Cost	228	(1)
Unsupervised Classification and Cluster Maps	228	(1)
A Clustering Example	228	(2)
A Single Pass Clustering Technique	230	(3)
Single Pass Algorithm	230	(1)
Advantages and Limitations	231	(1)
Strip Generation Parameter	232	(1)
Variations on the Single Pass Algorithm	232	(1)
An Example	232	(1)
Agglomerative Hierarchical Clustering	233	(2)
Clustering by Histogram Peak Selection	235	(4)
References for Chapter 9	237	(1)
Problems	237	(2)
Feature Reduction	239	(20)
Feature Reduction and Separability	239	(1)
Separability Measures for Multivariate Normal Spectral Class Models	239	(8)
Distribution Overlaps	239	(1)
Divergence	240	(1)
A General Expression	240	(2)
Divergence of a Pair of Normal Distributions	242	(1)
Use of Divergence for Feature Selection	242	(1)
A Problem with Divergence	243	(1)
The Jeffries-Matusita (JM) Distance	244	(1)
Definition	244	(1)
Comparison of Divergence and JM Distance	245	(1)
Transformed Divergence	245	(1)
Definition	245	(1)
Relation between Transformed Divergence and Probability of Correct Classification	246	(1)
Use of Transformed Divergence in Clustering	247	(1)
Separability Measures for Minimum Distance Classification	247	(1)
Feature Reduction by Data Transformation	247	(12)
Feature Reduction Using the Principal Components Transformation	247	(2)
Canonical Analysis as a Feature Selection Procedure	249	(1)
Within Class and Among Class Covariance Matrices	250	(1)
A Separability Measure	251	(1)
The Generalised Eigenvalue Equation	252	(1)
An Example	253	(2)
Arithmetic Transformations	255	(1)
References for Chapter 10	256	(1)
Problems	256	(3)
Image Classification Methodologies	259	(34)
Introduction	259	(1)
Supervised Classification	259	(3)
Outline	259	(1)
Determination of Training Data	260	(1)
Feature Selection	260	(1)
Detecting Multimodal Distributions	261	(1)
Presentation of Results	261	(1)
Effect of Resampling on Classification	262	(1)
Unsupervised Classification	262	(2)
Outline, and Comparison with Supervised Methods	262	(2)
Feature Selection	264	(1)
A Hybrid Supervised/Unsupervised Methodology	264	(1)
The Essential Steps	264	(1)
Choice of the Clustering Regions	265	(1)
Rationalisation of the Number of Spectral Classes	265	(1)
Assessment of Classification Accuracy	265	(5)
Case Study 1: Irrigated Area Determination	270	(6)
Background	270	(1)
The CSIRO-ORSER Image Analysis Software	270	(1)
The Study Region	270	(1)
Clustering	271	(3)
Signature Generation	274	(1)
Classification and Results	274	(1)
Concluding Remarks	274	(2)
Case Study 2: Multitemporal Monitoring of Bush Fires	276	(7)
Background	276	(1)
Simple Illustration of the Technique	276	(2)
The Study Area	278	(1)
Registration	278	(1)
Principal Components Transformation	279	(2)
Classification of Principal Components Imagery	281	(2)
Hierarchical Classification	283	(10)
The Decision Tree Classifier	283	(1)
Decision Tree Design	283	(3)
Progressive Two-Class Decision Classifier	286	(1)
Error Accumulation in a Decision Tree	287	(1)
Reference for Chapter 11	288	(2)
Problems	290	(3)
Data Fusion	293	(20)
The Stacked Vector Approach	293	(1)
Statistical Methods	294	(1)
The Theory of Evidence	295	(3)
The Concept of Evidential Mass	295	(1)
Combining Evidence -- the Orthogonal Sum	296	(2)
Decision Rule	298	(1)
Knowledge-Based Image Analysis	298	(15)
Knowledge Processing: Emulating Photointerpretation	299	(1)
Fundamentals of a Knowledge-Based Image Analysis System	300	(1)
Structure	300	(1)
Representation of Knowledge: Rules	301	(1)
The Inference Mechanism	302	(1)
Handling Multisource and Multisensor Data	303	(2)
An Example	305	(1)
Rules as Justifiers for a Labelling Proposition	306	(1)
Endorsement of a Labelling Proposition	307	(1)
Knowledge Base and Results	308	(2)
Reference for Chapter 12	310	(2)
Problems	312	(1)
Interpretation of Hyperspectral Image Data	313	(26)
Data Characteristics	313	(2)
The Challenge of Interpretation	315	(4)
Data Volume	315	(1)
Redundancy	315	(2)
The Need for Calibration	317	(1)
The Problem of Dimensionally: the Hughes Phenomenon	318	(1)
Data Calibration Techniques	319	(3)
Detailed Radiometric Correction	319	(1)
Data Normalisation	320	(1)
Approximate Radiometric Correction	321	(1)
Interpretation by Spectral Analysis	322	(5)
Spectroscopic Analysis	322	(1)
Spectral Angle Mapping	322	(1)
Library Searching Techniques	323	(1)
Binary Spectral Codes	323	(2)
Matching Algorithms	325	(2)
Hyperspectral Interpretation by Statistical Methods	327	(2)
Limitations of Traditional Thematic Mapping Procedures	327	(1)
Block-based Maximum Likelihood Classification	327	(2)
Feature Reduction	329	(2)
Feature Selection	329	(1)
Spectral Transformations	329	(2)
Feature Selection from Principal Components Transformed Data	331	(1)
Compression of Hyperspectral Data	331	(3)
Spectral Unmixing: End Member Analysis	334	(5)
Reference for Chapter 13	335	(1)
Problems	336	(3)
Appendix A -- Satellite Altitudes and Periods	339	(2)
References for Appendix A	340	(1)
Appendix B -- Binary Representation of Decimal Numbers	341	(2)
Appendix C -- Essential Results from Vector and Matrix Algebra	343	(6)
C.1 Definition of a Vector and a Matrix	343	(2)
C.2 Properties of Matrices	345	(1)
C.3 Multiplication, Addition and Subtraction of Matrices	346	(1)
C.4 The Eigenvalues and Eigenvectors of a Matrix	346	(1)
C.5 Some Important Matrix, Vector Operations	347	(1)
C.6 An Orthogonal Matrix -- the Concept of Matrix Transpose	347	(1)
C.7 Diagonalisation of a Matrix	347	(2)
References for Appendix C	348	(1)
Appendix D -- Some Fundamental Material from Probability and Statistics	349	(4)
D.1 Conditional Probability	349	(1)
D.2 The Normal Probability Distribution	350	(3)
D.2.1 The Univariate Case	350	(1)
D.2.2 The Multivariate Case	350	(1)
Reference for Appendix D	351	(2)
Appendix E -- Penalty Function Derivation of the Maximum Likelihood Decision Rule	353	(4)
E.1 Loss Functions and Conditional Average Loss	353	(1)
E.2 A particular Loss Function	354	(3)
References for Appendix E	355	(2)
Subject Index	357