Muutke küpsiste eelistusi

E-raamat: Spatial Analysis for Radar Remote Sensing of Tropical Forests [Taylor & Francis e-raamat]

,
  • Formaat: 335 pages, 28 Tables, black and white; 110 Illustrations, color
  • Sari: SAR Remote Sensing
  • Ilmumisaeg: 25-Mar-2021
  • Kirjastus: CRC Press
  • ISBN-13: 9780429290657
Teised raamatud teemal:
  • Taylor & Francis e-raamat
  • Hind: 216,96 €*
  • * hind, mis tagab piiramatu üheaegsete kasutajate arvuga ligipääsu piiramatuks ajaks
  • Tavahind: 309,94 €
  • Säästad 30%
Spatial Analysis for Radar Remote Sensing of Tropical ForestsSuurem pilt
Taylor & Francis e-raamatudRohkem infot Taylor & Francis e-raamatute kohta
Raamatu kodulehekülg: https://www.taylorfrancis.com/books/9780429290657

Spatial Analysis for Radar Remote Sensing of Tropical Forests is based on the authors’ extensive involvement in Synthetic Aperture Radar (SAR) mapping projects, targeting the health of an earth ecosystem with great relevance for climate change studies: the tropical forests. The subject is developed from a vantage point provided by analysis in a combined space, scale (frequency), time, wavelength, polarization domain.

The combination of space and scale offers the capability to zoom in and out like a virtual microscope to the resolution in tune with the underlying ecological phenomenon. It also enables statistical measures (correlations) related to the forest spatial distribution in case of backscatter, or to the canopy height variations in case of interferometric observations.

The time dimension brings into play measures of the ecosystem dynamics, such as the flooding extent in the swamp forests, deforestation or degradation events.

Wavelength and polarization agility extend the abovementioned capabilities by radar observations that are in tune with particular characteristics of the forest and terrain layers.

The book’s spotlight is on radar spatial random fields, these being populated by either backscatter observations or elevation data from interferometric SAR. The basic tenet here is that the spatial statistic of the fields measured by the wavelet variance (in stationary or non-stationary situations) carries fingerprints of the forest structure.

Features:

  • Uniquely focused on specific techniques that provide multi-resolution spatial and temporal analysis of forest structure characteristics and changes.

  • Examines several large and important international remote sensing projects aimed at documenting entire tropical ecosystems.

  • Provides novel wavelet methods for tropical forest structural measures.

As the first book on this topic, this composite approach appeals to both students learning through important case studies and to researchers finding new ideas for future studies.

Preface xi
Acknowledgements xv
The Authors xvii
List of Abbreviations
xix
List of Figures and Tables
xxiii
PART I Sarcheology: The Era of the Big Radar Mosaics
Chapter 1 The Dawn of the SAR Mosaics Era: The ESA-JRC Central Africa Mosaic Project
3(12)
1.1 Radar Mosaics: What and Why
3(5)
1.2 The CAMP Data Processing Machine
8(1)
1.3 Radiometry
9(4)
1.3.1 Radiometric Changes in Time
9(2)
1.3.2 Within Tile Radiometric Changes in range
11(1)
1.3.3 Quantization Noise
11(2)
References
13(2)
Chapter 2 The L-band Breed: The GRFM Africa Radar Mosaic
15(16)
2.1 The GRFM Project
15(2)
2.2 The GRFM Africa Processing Chain
17(2)
2.2.1 Input Datasets
17(1)
2.2.2 Data Flow
18(1)
2.3 Geolocation
19(6)
2.3.1 The Block Adjustment Method
19(5)
2.3.2 Geolocation Validation
24(1)
2.4 Wavelet Multiresolution Decomposition
25(1)
2.4.1 Multiresolution Products
26(1)
2.5 The GRFM Africa Mosaic Second Edition
26(2)
References
28(3)
Chapter 3 The GRFM-CAMP Thematic Products
31(18)
3.1 From Backscatter to a Thematic Map
31(1)
3.2 Vegetation Classes
32(2)
3.3 Map Compilation Methods
34(2)
3.4 Complementarity of Radar Sensors
36(3)
3.5 Validation
39(1)
3.6 Tour of Relevant Features
40(6)
References
46(3)
Chapter 4 Evolution of the Species: The ALOS PALSAR Africa Mosaic
49(24)
4.1 Introduction
49(2)
4.2 The Mosaic Processing Chain
51(1)
4.3 Correction of Range Dependent Radiometric Bias in Path Images
52(1)
4.4 Correction for Additive Thermal Noise in HV Strip Images
53(1)
4.5 Radiometric Inter-strip Mosaic Balancing
53(4)
4.6 Geocoding
57(1)
4.7 Radiometric Normalization for Topographic Effects
58(5)
4.7.1 Correction of Effective Scattering Area
58(2)
4.7.2 Correction for the Dependence of the Backscattering Coefficient on Incidence Angle
60(1)
4.7.3 Assessment of the Radiometric Correction for Topography
61(2)
4.8 Overview of the Thematic Information Content
63(6)
4.8.1 Comparison with the GRFM Africa Dataset
63(3)
4.8.2 Grass and Woody Savannas
66(1)
4.8.3 Flooded Forest
66(1)
4.8.4 Plantations
67(1)
4.8.5 Secondary Forest
67(2)
References
69(4)
PART II Measures of SAR Random Fields in the Scale--Space--Time Domain
Chapter 5 The Stuff Backscatter Random Fields Are Made Of
73(22)
5.1 Introduction
73(1)
5.2 Transport Theory
74(8)
5.2.1 An Illustrative Case: Propagation Through A Plane Parallel Medium
76(6)
5.3 The UTA Wave Scattering Model for Layered Vegetation
82(4)
5.4 Backscatter Simulation for a Dense Tropical Primary Rain Forest
86(6)
References
92(3)
Chapter 6 Statistical Measures of SAR Random Spatial Fields: Fingerprints of the Forest Structure
95(78)
6.1 Introduction
95(1)
6.2 Random Fields from Backscatter Observations
96(7)
6.3 Random Fields from InSAR Coherence Observations
103(1)
6.4 Wavelet Based Textural Measures of Random Fields
104(6)
6.5 Connection between Wavelet Space-Scale Analysis and Fourier Spectral Analysis
110(6)
6.5.1 White Noise
111(1)
6.5.2 1/f Process
112(1)
6.5.3 Correlated Surface (Gamma Distributed RCS) with Exponential ACF (Lorentzian Spectrum)
112(2)
6.5.4 Correlated Surface with Exponential Cosine ACF
114(1)
6.5.5 Effects from Coherent Imaging and Illumination Beam Size
114(1)
6.5.6 Cross-correlation between Two Stationary Processes with a Gaussian CCF
115(1)
6.6 Accuracy of Wavelet Variance Estimators
116(13)
6.6.1 Prelude: Probability Density Function of the Wavelet Coefficients of a Speckle Pattern
117(4)
6.6.2 Expected Value and Variance of the Wavelet Variance Estimator
121(1)
6.6.2.1 Uncorrelated Speckle Pattern
121(5)
6.6.2.2 Correlated Speckle
126(3)
6.7 Tools for Textural Analysis of SAR Random Fields
129(5)
6.7.1 A Multi-Voice Discrete Wavelet Transform
129(3)
6.7.2 Wavelet Signatures
132(2)
6.7.3 Wavelet Spectra
134(1)
6.8 WASS Analysis of SAR Backscatter Fields
134(9)
6.8.1 Lowland Rainforest and Swamp Forest Signatures in ERS-1 Data
134(3)
6.8.2 TanDEM-X Signatures in the same Thematic Context
137(3)
6.8.3 Intact and Degraded Forest Detection by Functional Analysis of WASS Signatures
140(3)
6.9 WASS Analysis of InSAR and LiDAR Digital Surface Models
143(6)
6.10 2D Wavelet Variance Spectra of Backscatter Fields: Toward a Textural Classifier
149(7)
6.10.1 A Test Case: Texture-Based Forest Mapping in the Congo Floodplain by ERS-1 Data
153(1)
6.10.2 Floodplain Mapping Revisited by Sentinel-1 data
154(1)
6.10.3 An (Experimental) Wavelet Spectrum Functional Classifier
155(1)
6.11 Extension to Polarimetry
156(15)
6.11.1 The WASP of Correlated Backscatter Textures: A Numerical Model
157(10)
6.11.2 WASP Analysis of a PALSAR Full-Pol Data Set
167(4)
References
171(2)
Chapter 7 Hitting Corners: The Lipschitz Regularity, a Measure of Discontinuities in Radar Images Connected with Forest Spatial Distribution
173(64)
7.1 Introduction
173(1)
7.2 The Lipschitz Condition
173(3)
7.3 Singular Functions and Lip Parameters Estimated by Wavelet Maxima Trajectories in the Scale Domain
176(18)
7.3.1 Step Function
177(2)
7.3.2 Cusp
179(3)
7.3.3 Impulse
182(3)
7.3.4 Smoothed Singularity
185(3)
7.3.5 Non-Isolated Singularities
188(3)
7.3.6 Effect of Speckle
191(3)
7.4 A Monte Carlo Simulator of Polarimetric SAR Backscatter Discontinuities
194(3)
7.5 Experiments using Simulated Signals
197(20)
7.5.1 Toy Signals with Simple Discontinuities
197(6)
7.5.2 Margin between a Clear-Cut and a Dense Forest
203(9)
7.5.3 Edge on Tilted Terrain
212(5)
7.6 Lipschitz Regularity in Real SAR Data
217(16)
7.6.1 TanDEM-X Backscatter Data
217(10)
7.6.2 TanDEM-X Coherence Data
227(6)
7.7 Image-Wide Representations of Lipschitz Parameters
233(2)
References
235(2)
Chapter 8 The Beauty Farm: A Wavelet Method for Edge Preserving Piece-wise Smooth Approximations of Radar Images
237(18)
8.1 The Image Model and a Conceptual View of the Method
238(1)
8.2 The Computational Engine
239(2)
8.3 Problems Related to Multiplicative Speckle Noise
241(5)
8.4 Issues Related to Textural Edges
246(1)
8.5 Maxima Linking
247(1)
8.6 From Theory to Practice: A Tropical Forest Cover Mapping Exercise Using Smooth Approximations of GRFM SAR Data
247(6)
8.6.1 Processing Methods
248(1)
8.6.1.1 Region Growing
248(1)
8.6.1.2 NMP Classifier
249(1)
8.6.2 Test Sites and Thematic Class Definition
249(1)
8.6.3 Selected Results
250(3)
References
253(2)
Chapter 9 The Cleaning Service: A Multi-temporal InSAR Coherence Magnitude Filter
255(24)
9.1 Rationale
255(3)
9.2 The Filter Machinery
258(2)
9.3 Generation of a Testing Dataset
260(3)
9.4 Test Cases using TanDEM-X Data
263(11)
9.5 Temporal Features
274(3)
References
277(2)
Chapter 10 Proxies of Forest Volume Loss And Gain by Differencing InSAR DSMs: Fingerprints of Forest Disturbance
279(36)
10.1 Motivation
279(2)
10.2 Study site
281(1)
10.3 TanDEM-X Data
282(1)
10.4 Methods
283(9)
10.4.1 DSM Difference Data Set Generation and Calibration
283(1)
10.4.2 Object-Based Change Detection
284(1)
10.4.3 Change Objects Refinement
285(1)
10.4.4 Variance of the Within-Object Mean Height Difference Estimator
285(1)
10.4.5 Effect Size
286(1)
10.4.6 Probability of object detection by statistical decision theory
287(1)
10.4.6.1 Neyman--Pearson approach
288(1)
10.4.6.2 Bayesian Approach
289(1)
10.4.7 Object Shape
290(1)
10.4.8 Characterization of Objects by Contextual Information
291(1)
10.4.8.1 Distance from Roads
291(1)
10.4.8.2 Attributes by Land Management
291(1)
10.5 Factors Influencing the DSM Change Magnitude
292(3)
10.5.1 Forest Vertical Structure and Spatial Distribution (Forest Density)
292(1)
10.5.2 Environmental Conditions (Seasonality and Rainfall)
292(1)
10.5.3 Dependence on Instrument Parameters
292(1)
10.5.3.1 Volume Only
293(1)
10.5.3.2 Volume over Ground
294(1)
10.6 Analysis
295(14)
10.6.1 ADSM Magnitude and Area Descriptive Statistic
295(1)
10.6.2 Standard Error of the Object Mean
296(4)
10.6.3 Effect Size
300(1)
10.6.4 Object Detection by Statistical Decision Theory
301(1)
10.6.5 Spatial Location of Objects
302(1)
10.6.6 Objects' Proximity to Roads
303(1)
10.6.7 Change in Objects by Land Management
303(1)
10.6.8 Shape Analysis
304(1)
10.6.8.1 Fractal Exponent
304(2)
10.6.8.2 Rectangularity
306(2)
10.6.8.3 Regular Boundary Shapes in Land Management Units
308(1)
10.7 Comparison between Objects Detected by InSAR ΔDSM and by Optical Imagery
309(1)
10.8 Concluding Remarks
310(1)
References
311(4)
Appendix 315(20)
Index 335
Gianfranco (Frank) De Grandi received a doctorate degree in Physics Engineering with honors from the Politecnico Milano, Italy, in 1973. He later joined the European Commission Joint Research Center, Ispra, Italy, where he has performed research in signal processing for application areas such as gamma ray spectroscopy, data communications, and radar remote sensing. From 1986 to 1989 he headed the signal processing section of the Electronics Division, JRC, where he introduced VLSI design technology and conducted research, in cooperation with Bellcore, on packet video, and in cooperation with ITALTEL Italy on the European digital mobile phone network. In 1989 he joined the JRC Institute for Remote Sensing Applications, where he started his research activity in radar remote sensing for earth observations, and in particular in SAR polarimetry. From 1997 to 2001 he has served as assistant professor with the Faculte' de Feresterie et Geomatique, Universite' Laval, Quebec, PQ, Canada. He retired from active service at the European Commission in 2012. He continued through 2017 his activity in science and education as a visiting scholar at the University of Wales at Aberystwyth, Ceredigion, UK. Dr. De Grandi collaborated with several laboratories in the US, such as Lawrence Livermore National Lab, and Los Alamos National Lab for nuclear safeguards, Caltech Jet Propulsion Laboratory and the US Naval Research Lab for radar remote sensing. At the EC Joint Research Center (JRC) he was team leader of the Central Africa Mosaic Project (CAMP), a joint initiative of the JRC and the European Space Agency (ESA), which for the first time produced a regional scale high resolution radar map of Central Africa. He was principal investigator with the Japan Aerospace Exploration Agency (JAXA) Global Rain Forest Mapping and Global Boreal Forest Mapping projects, of the JAXA EORC Kyoto & Carbon Initiative for ALOS PALSAR. He was an international collaborator in a NASA Carbon Cycle Science project, led by U. Washington, WA USA. In 2002 Dr. De Grandi was elected IEEE Fellow for his seminal work in continental scale vegetation mapping using high resolution SAR mosaics, and innovative contributions in the area of information extraction from SAR data. He is a member of the IEEE Geoscience and Remote Sensing society, and the IEEE Signal Processing Society.



Elsa Carla De Grandi received a B.Sc. degree in Physical Geography in 2011 and a M.Sc. degree in 2012 (with distinction) in Remote Sensing and Geography from Aberystwyth University, Aberystwyth, UK. In 2017 she was awarded a PhD in Remote Sensing (Atmospheric & Environmental Science) from the University of Edinburgh. Her PhD focused on developing and testing novel SAR and InSAR methods for mapping deforestation and forest degradation in tropical forests. Since 2019 Elsa has been working as an Earth Observation Engineer at GMV NSL in UK. She is also acting as project manager for several projects and contributing to proposal writing particularly on the use of SAR for environmental monitoring. She was a reviewer for IEEE Transactions on Geoscience and Remote Sensing, has published peer-reviewed papers in international journals, and delivered talks on radar remote sensing of tropical forest at several conferences.
Püsilink: https://www.kriso.ee/db/9780429290657_pe.html
Märksõnad: