Advances in Hyperspectral Image Processing Techniques E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Serie: Wiley - IEEE
• Sprache: Englisch

Seitenzahl: 1131

Veröffentlichungsjahr: 2022

Beschreibung

Advances in Hyperspectral Image Processing Techniques Authoritative and comprehensive resource covering recent hyperspectral imaging techniques from theory to applications Advances in Hyperspectral Image Processing Techniques is derived from recent developments of hyperspectral imaging (HSI) techniques along with new applications in the field, covering many new ideas that have been explored and have led to various new directions in the past few years. The work gathers an array of disparate research into one resource and explores its numerous applications across a wide variety of disciplinary areas. In particular, it includes an introductory chapter on fundamentals of HSI and a chapter on extensive use of HSI techniques in satellite on-orbit and on-board processing to aid readers involved in these specific fields. The book's content is based on the expertise of invited scholars and is categorized into six parts. Part I provides general theory. Part II presents various Band Selection techniques for Hyperspectral Images. Part III reviews recent developments on Compressive Sensing for Hyperspectral Imaging. Part IV includes Fusion of Hyperspectral Images. Part V covers Hyperspectral Data Unmixing. Part VI offers different views on Hyperspectral Image Classification. Specific sample topics covered in Advances in Hyperspectral Image Processing Techniques include: * Two fundamental principles of hyperspectral imaging * Constrained band selection for hyperspectral imaging and class information-based band selection for hyperspectral image classification * Restricted entropy and spectrum properties for hyperspectral imaging and endmember finding in compressively sensed band domain * Hyperspectral and LIDAR data fusion, fusion of band selection methods for hyperspectral imaging, and fusion using multi-dimensional information * Advances in spectral unmixing of hyperspectral data and fully constrained least squares linear spectral mixture analysis * Sparse representation-based hyperspectral image classification; collaborative hyperspectral image classification; class-feature weighted hyperspectral image classification; target detection approach to hyperspectral image classification With many applications beyond traditional remote sensing, ranging from defense and intelligence, to agriculture, to forestry, to environmental monitoring, to food safety and inspection, to medical imaging, Advances in Hyperspectral Image Processing Techniques is an essential resource on the topic for industry professionals, researchers, academics, and graduate students working in the field.

Leseprobe

ADVANCES IN HYPERSPECTRAL IMAGE PROCESSING TECHNIQUES

Edited by

Chein‐I Chang

Remote Sensing Signal and Image Processing Laboratory, Department of Computer Science and Electrical Engineering, University of Maryland Baltimore County, Baltimore, MD, USA

This edition first published 2023© 2023 John Wiley & Sons, Inc.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Chein‐I Chang to be identified as the author of the editorial material in this work has been asserted in accordance with law.

Registered OfficeJohn Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA

Editorial Office111 River Street, Hoboken, NJ 07030, USA

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of WarrantyIn view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging‐in‐Publication DataName: Chang, Chein-I, editor.Title: Advances in hyperspectral image processing techniques / edited by Chein-I Chang.Description: Hoboken, NJ : Wiley, 2023. | Includes bibliographical references and index.Identifiers: LCCN 2022022693 (print) | LCCN 2022022694 (ebook) | ISBN 9781119687764 (cloth) | ISBN 9781119687757 (adobe pdf) | ISBN 9781119687771 (epub)Subjects: LCSH: Hyperspectral imaging.Classification: LCC TR267.73 .A38 2023 (print) | LCC TR267.73 (ebook) | DDC 771–dc23/eng/20220829LC record available at https://lccn.loc.gov/2022022693LC ebook record available at https://lccn.loc.gov/2022022694

Cover Design: WileyCover Image: © Shutterstock/ASVMAGZ

Editor Biography

Chien-I Chang received his Ph.D. degree in electrical engineering from the University of Maryland, College Park, and has been with the University of Maryland Baltimore County (UMBC), since 1987 where he is a professor in the Department of Computer Science and Electrical Engineering. Dr. Chang authored four books: Hyperspectral Imaging: Techniques for Spectral Detection and Classification published by Kluwer Academic Publishers in 2003; Hyperspectral Data Processing: Algorithm Design and Analysis, John Wiley & Sons, 2013; Real Time Progressive Hyperspectral Image Processing: Endmember Finding and Anomaly Detection in 2016 by Springer; and Recursive Hyperspectral Sample and Band Processing: Algorithm Architecture and Implementation, Springer 2017. In addition, he has edited two books, Recent Advances in Hyperspectral Signal and Image Processing, 2006, and Hyperspectral Data Exploitation: Theory and Applications, and co-edited with A. Plaza a book on High Performance Computing in Remote Sensing, CRC Press, 2007. Dr. Chang is a life fellow of IEEE and a fellow of SPIE.

List of Contributors

Adam BekitRemote Sensing Signal and Image Processing Laboratory, Department of Computer Science and Electrical Engineering, University of Maryland Baltimore County, Baltimore, MD, USA

Chein‐I ChangCenter for Hyperspectral Imaging in Remote Sensing (CHIRS), Information and Technology College, Dalian Maritime University, Dalian, ChinaRemote Sensing Signal and Image Processing Laboratory, Department of Computer Science and Electrical Engineering, University of Maryland Baltimore County, Baltimore, MD, USA

Shuhan ChenDepartment of Electrical Engineering, Zhejiang University, Hangzhou, China

Charles J. Della‐PortaRemote Sensing Signal and Image Processing Laboratory, Department of Computer Science and Electrical Engineering, University of Maryland Baltimore County, Baltimore, MD, USA

Qian DuDepartment of Electrical and Computer Engineering, Mississippi State University, Starkville, MS, USA

Lianru GaoKey Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China

Chiru GeSchool of Information Science and Engineering, Shandong Normal University, Jinan, China

Zhu HanKey Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaandCollege of Resources and Environment, University of Chinese Academy of Sciences, Beijing, China

Risheng HuangCollege of Electrical Engineering, Zhejiang University, Hangzhou, Zhejiang, China

Bernard LampeRemote Sensing Signal and Image Processing Laboratory, Department of Computer Science and Electrical Engineering, University of Maryland Baltimore County, Baltimore, MD, USA

Jie LeiState Key Laboratory of Integrated Services Networks, Xidian University, Xi’an, China

Hsiao‐Chi LiDepartment of Electrical Engineering, National Taipei University of Technology, Taipei, Taiwan, Republic of China

Jiaojiao LiState Key Laboratory of Integrated Services Networks, School of Telecommunications Engineering, Xidian University, Xian, China

Jun LiGuangdong Provincial Key Laboratory of Urbanization and Geo‐Simulation, School of Geography and Planning, Sun Yat‐sen University, Guangzhou, China

Tong LiSatellite Application Division, Shanghai Institute of Satellite Engineering, Shanghai, China

Wei LiCollege of Information and Electronics, Beijing Institute of Technology, Beijing, ChinaandSchool of Information and Electronics, Beijing Institute of Technology, Beijing, China

Xiaorun LiCollege of Electrical Engineering, Zhejiang University, Hangzhou, Zhejiang, China

Yunsong LiState Key Laboratory of Integrated Services Networks, Xidian University, Xi’an, China

Kai LiuState Key Laboratory of Integrated Services Networks, Xidian University, Xi’an, China

Xiaochen LuSchool of Information Science and Technology, Donghua University, Shanghai, China

Wenfei LuoSchool of Geographical Science, South China Normal University, Guangzhou, China

Mingyuan PengLand Satellite Remote Sensing Application Center, MNR, Beijing, China

Shen‐En QianCanadian Space Agency, Saint‐Hubert, Quebec, Canada

Yuntao QianCollege of Computer Science, Zhejiang University, Hangzhou, China

Xiaodi ShangCenter for Hyperspectral Imaging in Remote Sensing (CHIRS), School of Information and Technology, Dalian Maritime University, Dalian, China

Meiping SongCenter for Hyperspectral Imaging in Remote Sensing (CHIRS), Information and Technology College, Dalian Maritime University, Dalian, China

Xu SunKey Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China

Xuejian SunHyperspectral Remote Sensing Application Division, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China

Keyan WangState Key Laboratory of Integrated Services Networks, Xidian University, Xi’an, China

Lin WangSchool of Physics and Optoelectronic Engineering, Xidian University, Xi'an, China

Yulei WangCenter for Hyperspectral Imaging in Remote Sensing (CHIRS), School of Information and Technology, Dalian Maritime University, Dalian, China

Weiying XieState Key Laboratory of Integrated Services Networks, Xidian University, Xi’an, China

Fengchao XiongSchool of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, China

Bai XueDepartment of Computer Science and Electrical Engineering, University of Maryland Baltimore County, Baltimore, MD, USA

Minchao YeCollege of Information Engineering, China Jiliang University, Hangzhou, China

Chunyan YuCenter for Hyperspectral Imaging in Remote Sensing (CHIRS), Information and Technology College, Dalian Maritime University, Dalian, China

Haoyang YuCenter for Hyperspectral Imaging in Remote Sensing (CHIRS), Information Science and Technology College, Dalian Maritime University, Dalian, China

Bing ZhangKey Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaandCollege of Resources and Environment, University of Chinese Academy of Sciences, Beijing, China

Junping ZhangSchool of Electronics and Information Engineering, Harbin Institute of Technology, Harbin, China

Lifu ZhangHyperspectral Remote Sensing Application Division, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China

Xia ZhangHyperspectral Remote Sensing Application Division, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, China

Liaoying ZhaoSchool of Computer Science, Hangzhou Dianzi University, Hangzhou, Zhejiang, China

Xiaoyang ZhaoHyperspectral Remote Sensing Application Division, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaandCollege of Natural Resources and Environment, University of Chinese Academy of Sciences, Beijing, China

Shengwei ZhongSchool of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing, China

Jun ZhouSchool of Information and Communication Technology, Griffith University, Nathan, QLD, Australia

Lina ZhuangDepartment of Mathematics, Hong Kong Baptist University, Hong Kong, China

Preface

Hyperspectral imaging (HSI) has been around more than three decades and also become a mature advanced remote sensing technology in many applications ranging from defense and intelligence, agriculture, environmental monitoring to food inspection and safety, and medical imaging. Since my first book, Hyperspectral Imaging: Techniques for Spectral Detection and Classification, was published by Kluwer Academic/Plenum Publishers in 2003, nearly 20 years have been passed. It has also been my 30 years of research devoted to HSI that has evidenced a rapid growth in remote sensing over the past years with tens of thousands of reports published every year in this fascinating area.

This book entitled “Advances in Hyperspectral Image Processing Techniques” does not intend to cover all the areas in hyperspectral imaging but rather addresses several key areas of interest in HSI based on my most recent research carried out in the Remote Sensing Signal and Image Processing Laboratory (RSSIPL) at the University of Maryland Baltimore County (UMBC) and the Center of Hyperspectral Imaging in Remote Sensing (CHIRS), Dalian Maritime University, Dalian, China. As a result, the topics of many chapters in this book reflect this nature and are new results that have not been discussed and covered in any of my previous books. To make this book a valuable and useful reference, all contributors are experts in their own areas and were invited to write book chapters sharing their research experiences and findings. In particular, all the contributors are either my long‐time close friends, my colleagues in CHIRS, my former PhD students, post‐doctors, visiting scholars, or exchanging PhD students. To this end, this book simply represents a very small portion of research in HSI and never tries to cover a full range of HSI areas.

Twenty chapters are included in this book and written by many renowned scholars and researchers in HSI who were cordially invited to make contributions. Many of them have established reputation in their own fields and have been kind to agree to write their own book chapters. All the book chapters can be categorized in six parts – Part I: General Theory, Part II: Band Selection for Hyperspectral Imaging, Part III: Compressive Sensing for Hyperspectral Imaging, Part IV: Fusion of Hyperspectral Imaging, Part V: Hyperspectral Data Unmixing, and Part VI: Hyperspectral Image Classification – each of which is summarized as follows.

Part I: General Theory

Chapter 1

:

Introduction: Two Fundamental Principles Behind Hyperspectral Imaging Chein‐I Chang

This chapter explores two fundamental principles: pigeon‐hole principle and orthogonality principle behind design and development of hyperspectral image processing algorithms in various applications such as pigeon‐hole principle that finds its applications in determining the number of endmembers needed for spectral unmixing, spectrally distinct signatures by virtual dimensionality, orders of low rank and sparse representation/low rank and sparse matrix decomposition, and the number of bands to be selected for band selection and orthogonality principle which can be used as a criterion to derive the orthogonal subspace projection (OSP) technique in developing algorithms for finding endmembers, detecting subpixel targets, anomalies as well as OSP‐based low rank and sparse matrix decomposition in target/anomaly detection, and OSP‐based detection theory applied to classification.

Chapter 2

:

Overview of Hyperspectral Imaging Remote Sensing from Satellites Shen‐En Qian

This chapter offers a valuable overview of design and development hyperspectral imaging sensors currently operated in space and their future developments for space exploration.

Chapter 3

:

Efficient Hardware Implementation for Hyperspectral Anomaly and Target Detection Jie Lei, Weiying Xie, Jiaojiao Li, Keyan Wang, Kai Liu, and Yunsong Li

This chapter introduces some real‐time detectors including deep pipelined background statistics (DPBS)‐constrained energy minimization (CEM), deep brief network (DBN)‐Reed‐Xiaoli anomaly detector (RXD), Fast‐automatic target‐generation process (ATGP), and fast‐morphological reconstruction and the simplified guided filtering detector (MGD), implemented on FPGA considering the concept of model‐based design, which overcomes the drawbacks of high difficulty and long period in the development of traditional hardware systems.

Part II: Band Selection for Hyperspectral Imaging

Chapter 4

:

Constrained Band Selection for Hyperspectral Imaging Chein‐I Chang

This chapter reviews a particular type of band selection methods, constrained band selection that has two versions, band‐constrained band selection (BCBS) and target‐constrained band selection (TCBS), both of which are derived from the well‐known subpixel target detection, constrained energy minimization (CEM), and linearly constrained minimum variance (LCMV) filter.

Chapter 5

:

Band Subset Selection for Hyperspectral Imaging Chein‐I Chang

This chapter extends band selection (BS) in

Chapter 4

to band subset selection (BSS) in the sense that multiple bands are selected as a band subset rather than band by band selected sequentially as band selection does, particularly, single band‐constrained BS (BCBS) to multiple band‐constrained BSS (MBC‐BSS) and single‐target‐constrained BS (TCBS) to multiple‐target‐constrained BSS (MTC‐BSS).

Chapter 6

:

Progressive Band Selection Processing for Hyperspectral Image Classification Chunyan Yu, Meiping Song, and Chein‐I Chang

This chapter develops progressive band selection processing (PBSP) of hyperspectral image classification, which performs image classification stage‐by‐stage progressively in the sense that each stage performs HSIC by band subsets according to a

p

‐ary Huffman coding tree constructed by class classification priority (CCP).

Part III: Compressive Sensing for Hyperspectral Imaging

Chapter 7

:

Restricted Entropy and Spectrum Properties for Hyperspectral Imaging Chein‐I Chang and Bernard Lampe

This chapter derives two new compressive sensing properties for hyperspectral imaging, restricted entropy property (REP) and restricted spectrum property (RSP), which can be shown to preserve the entropy of a hyperspectral signature vector and the spectral similarity between two hyperspectral signatures in both the original data space and compressively sensed band domain so that REP and RSP preserve exploitation algorithm performance without the need for decompression so as to avoid specifying a sparse basis.

Chapter 8

:

Endmember Finding in Compressively Sensed Band Domain Chein‐I Chang and Adam Bekit

This chapter presents a compressive sensing approach to N‐FINDR that can find a

p

‐vertex simplex with the maximal simplex volume found by Sequential/Successive N‐FINDR with the restricted simplex volume property being proved to be preserved in compressively sensed band domain (CSBD).

Chapter 9

:

Hyperspectral Image Classification in Compressively Sensed Band Domain Charles J. Della‐Porta and Chein‐I Chang

This chapter explores the viability of performing classification for hyperspectral data via compressive sensing where a mathematical analysis is derived to show that the classification error can be expressed in terms of the restricted isometry constant so that the hyperspectral image classification in the original data space can be achieved in compressively sensed band domain (CSBD).

Part IV: Fusion for Hyperspectral Imaging

Chapter 10

:

Hyperspectral and LiDAR Data Fusion Qian Du, Wei Li, and Chiru Ge

This chapter reviews three deep learning based approaches to fusion of HSI and LiDAR data, Two‐Branch CNN for Joint Classification, Hierarchical Random Walk Network (HRWN), and Residual Network‐Based Probability Reconstruction Fusion (RNPRF), and conducts experiments to evaluate their comparative performance in land cover classification.

Chapter 11

:

Hyperspectral Data Fusion Using Multi‐Dimensional Information Lifu Zhang, Xia Zhang, Mingyuan Peng, Xuejian Sun, and Xiaoyang Zhao

This chapter selects six simple and easy‐to‐promote hyperspectral fusion methods to fuse multiple domestic satellite multispectral data with GF‐5 hyperspectral data and conducts a comprehensive study on their advantages and disadvantages based on five classic evaluation indicators, classification application accuracy, and running time cost.

Chapter 12

:

Fusion of Band Selection Methods for Hyperspectral Imaging Yulei Wang, Lin Wang, and Chein‐I Chang

This chapter develops an approach for fusing different band selection methods where two versions are derived, called progressive band selection fusion (PBSF) and simultaneous band selection fusion (SBSF), to fuse multiple band subsets selected by different band selection methods.

Part V: Hyperspectral Data Unmixing

Chapter 13

:

Model‐Inspired Deep Neural Networks for Hyperspectral Unmixing Yuntao Qian, Fengchao Xiong, Minchao Ye, and Jun Zhou

This chapter introduces several model‐inspired learning‐based methods derived from different spectral unmixing models and their corresponding optimization algorithms with the prior knowledge of the unmixing model and optimization algorithm incorporated into the network architectures.

Chapter 14

:

Analytical Fully Constrained Least Squares Linear Spectral Mixture Analysis Chein‐I Chang and Hsiao‐Chi Li

This chapter derives an analytical spectral unmixing technique, to be called analytical FCLS (AFCLS) that can find fully constrained least squares (FCLS) solutions in closed forms by Cramer’s rule analytically so that the AFCLS‐unmixed results using analytical solutions are more accurate than FCLS‐unmixed resulting from numerical solutions.

Chapter 15

:

Swarm Intelligence Optimization‐Based Spectral Unmixing Lianru Gao, Xu Sun, Zhu Han, Lina Zhuang, Wenfei Luo, and Bing Zhang

This chapter develops swarm intelligence algorithms that can be implemented in three spectral mixing models, linear mixing model (LMM), normal compositional model (NCM), and non‐linear mixing model (NLMM).

Chapter 16

:

Spectral‐Spatial Robust Nonnegative Matrix Factorization for Hyperspectral Unmixing Risheng Huang, Xiaorun Li, and Liaoying Zhao

This chapter develops a robust NMF using

l

1,2

norm (

l

1,2

‐RNMF) and further proposes a spectral‐spatial robust NMF model (SSRNMF) by incorporating

l

2,1

norm and

l

1,2

norm to take advantage of both spatial dimension and spectral dimension into consideration and achieves robustness to band noise and pixel noise simultaneously.

Part VI: Hyperspectral Image Classification

Chapter 17

:

Sparse Representation‐Based Hyperspectral Image Classification Haoyang Yu, Jun Li, Wei Li, and Bing Zhang

This chapter conducts studies and proposes four new classification methods, locality‐preserving sparse representation‐based classification (LPSRC), locality‐sensitive discriminant analysis‐group sparse representation‐based classification (LSDA‐GSRC), activity degree‐driven representation‐based classification (ADRC), and neighborhood ADRC (NADRC) for sparse representation‐based hyperspectral image classification.

Chapter 18

:

Collaborative Classification Based on Hyperspectral Images Junping Zhang, Xiaochen Lu, and Tong Li

This chapter focuses on the hyperspectral‐centered multi‐source image collaborative classification methods and analyzes in depth the probable problems and challenges in multi‐source image synergetic processing by fully exploring the complementary information of multi‐source images.

Chapter 19

:

Class Feature‐Weighted Hyperspectral Image Classification Shengwei Zhong, Jiaojiao Li, Xiaodi Shang, Shuhan Chen, and Chein‐I Chang

This chapter extends the commonly used hyperspectral image classification (HSIC) to class feature weighted HSIC where traditional classification measures, overall accuracy, and average accuracy are considered as special cases of several new introduced class feature measures to deal with imbalanced class and background issues.

Chapter 20

:

Target Detection Approaches to Hyperspectral Image Classification Chein‐I Chang, Bai Xue, and Chunyan Yu

This chapter shows that hyperspectral image classification can be in fact indeed interpreted by statistical detection theory to address background and imbalanced class issues where a 3D receiver operating characteristics (3D ROC) analysis is further used as evaluation tool for performance evaluation.

The ultimate goal of this book is to offer readers with a peek of the cutting‐edge research in HSI. In particular, it is my belief that this book can provide a useful guide and assistance for practitioners and engineers who are interested in HSI. Hopefully, the chapters presented in this book have just achieved what this book would like to accomplish.

This book was originally planned to invite my close friends and colleagues, former PhD students, post‐doctors, visiting scholars, and visiting exchanging PhD students to contribute to the book chapters to celebrate my seventieth birthday in 2020. It was scheduled to be published in 2020. Unfortunately, due to unexpected pandemic of COVID‐19, the original schedule has been pushed back to 2022.

Last but not least, I would like to express my sincere gratitude to all the contributors for their efforts in providing their book chapters to complete this book project. There is no doubt that this book cannot be completed without their dedicated involvement.

Chein‐I Chang

Center of Hyperspectral Imaging in Remote Sensing (CHIRS)

Information and Technology College

Dalian Maritime University

Dalian, China

Remote Sensing Signal and Image Processing Laboratory

Department of Computer Science and Electrical Engineering

University of Maryland Baltimore County

Baltimore, MD, USA

December 2021

Part IGeneral Theory