Techniques and Applications of Hyperspectral Image Analysis Techniques and Applications of Hyperspectral Image Analysis Edited by H F Grahn and P Geladi © 2007 John Wiley & Sons, Ltd ISBN: 978-0-470-01086-0 Techniques and Applications of Hyperspectral Image Analysis Hans F Grahn and Paul Geladi Copyright © 2007 John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England Telephone (+44) 1243 779777 Email (for orders and customer service enquiries): cs-books@wiley.co.uk Visit our Home Page on www.wiley.com 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, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing of the Publisher Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to permreq@wiley.co.uk, or faxed to (+44) 1243 770620 Designations used by companies to distinguish their products are often claimed as trademarks All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners The Publisher is not associated with any product or vendor mentioned in this book This publication is designed to provide accurate and authoritative information in regard to the subject matter covered It is sold on the understanding that the Publisher is not engaged in rendering professional services If professional advice or other expert assistance is required, the services of a competent professional should be sought The Publisher and the Author 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 fitness for a particular purpose The advice and strategies contained herein may not be suitable for every situation In 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 The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read No warranty may be created or extended by any promotional statements for this work Neither the Publisher nor the Author shall be liable for any damages arising herefrom Other Wiley Editorial Offices John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 6045 Freemont Blvd, Mississauga, Ontario, L5R 4J3, Canada Wiley also publishes its books in a variety of electronic formats Some content that appears in print may not be available in electronic books Anniversary Logo Design: Richard J Pacifico Library of Congress Cataloging in Publication Data Techniques and applications of hyperspectral image analysis / [edited by] Hans Grahn and Paul Geladi p cm Includes bibliographical references ISBN 978-0-470-01086-0 (cloth) Image processing—Statistical methods Multivariate analysis Multispectral photography I Grahn, Hans II Geladi, Paul TA1637.T42 2007 2007021097 621.36 7—dc22 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 978-0-470-01086-0 Typeset in 10/12pt Times by Integra Software Services Pvt Ltd, Pondicherry, India Printed and bound in Great Britain by TJ International, Padstow, Cornwall This book is printed on acid-free paper responsibly manufactured from sustainable forestry in which at least two trees are planted for each one used for paper production Contents Preface xiii List of Contributors xvii List of Abbreviations xix Multivariate Images, Hyperspectral Imaging: Background and Equipment Paul L M Geladi, Hans F Grahn and James E Burger 1.1 Introduction 1.2 Digital Images, Multivariate Images and Hyperspectral Images 1.3 Hyperspectral Image Generation 1.3.1 Introduction 1.3.2 Point Scanning Imaging 1.3.3 Line Scanning Imaging 1.3.4 Focal Plane Scanning Imaging 1.4 Essentials of Image Analysis Connecting Scene and Variable Spaces References Principles of Multivariate Image Analysis (MIA) in Remote Sensing, Technology and Industry Kim H Esbensen and Thorbjørn T Lied 2.1 Introduction 2.1.1 MIA Approach: Synopsis 2.2 Dataset Presentation 2.2.1 Master Dataset: Rationale 2.2.2 Montmorency Forest, Quebec, Canada: Forestry Background 1 5 14 17 17 18 18 18 19 vi CONTENTS 2.3 Tools in MIA 2.3.1 MIA Score Space Starting Point 2.3.2 Colour-slice Contouring in Score Cross-plots: a 3-D Histogram 2.3.3 Brushing: Relating Different Score Cross-plots 2.3.4 Joint Normal Distribution (or Not) 2.3.5 Local Models/Local Modelling: a Link to Classification 2.4 MIA Analysis Concept: Master Dataset Illustrations 2.4.1 A New Topographic Map Analogy 2.4.2 MIA Topographic Score Space Delineation of Single Classes 2.4.3 MIA Delineation of End-member Mixing Classes 2.4.4 Which to Use? When? How? 2.4.5 Scene-space Sampling in Score Space 2.5 Conclusions References Clustering and Classification in Multispectral Imaging for Quality Inspection of Postharvest Products Jacco C Noordam and Willie H A M van den Broek 3.1 Introduction to Multispectral Imaging in Agriculture 3.1.1 Measuring Quality 3.1.2 Spectral Imaging in Agriculture 3.2 Unsupervised Classification of Multispectral Images 3.2.1 Unsupervised Classification with FCM 3.2.2 FCM Clustering 3.2.3 cFCM Clustering 3.2.4 csiFCM 3.2.5 Combining Spectral and Spatial Information 3.2.6 sgFCM Clustering 3.3 Supervised Classification of Multispectral Images 3.3.1 Multivariate Image Analysis for Training Set Selection 3.3.2 FEMOS 3.3.3 Experiment with a Multispectral Image of Pine and Spruce Wood 3.3.4 Clustering with FEMOS Procedure 21 21 24 24 26 27 28 28 31 33 38 39 40 41 43 43 43 44 46 46 47 48 49 51 52 54 55 57 58 60 CONTENTS 3.4 Visualization and Coloring of Segmented Images and Graphs: Class Coloring 3.5 Conclusions References Self-modeling Image Analysis with SIMPLISMA Willem Windig, Sharon Markel and Patrick M Thompson 4.1 Introduction 4.2 Materials and Methods 4.2.1 FTIR Microscopy 4.2.2 SIMS Imaging of a Mixture of Palmitic and Stearic Acids on Aluminum foil 4.2.3 Data Analysis 4.3 Theory 4.4 Results and Discussion 4.4.1 FTIR Microscopy Transmission Data of a Polymer Laminate 4.4.2 FTIR Reflectance Data of a Mixture of Aspirin and Sugar 4.4.3 SIMS Imaging of a Mixture of Palmitic and Stearic Acids on Aluminum Foil 4.5 Conclusions References Multivariate Analysis of Spectral Images Composed of Count Data Michael R Keenan 5.1 5.2 5.3 5.4 Introduction Example Datasets and Simulations Component Analysis Orthogonal Matrix Factorization 5.4.1 PCA and Related Methods 5.4.2 PCA of Arbitrary Factor Models 5.4.3 Maximum Likelihood PCA (MLPCA) 5.4.4 Weighted PCA (WPCA) 5.4.5 Principal Factor Analysis (PFA) 5.4.6 Selecting the Number of Components 5.5 Maximum Likelihood Based Approaches 5.5.1 Poisson Non-negative Matrix Factorization (PNNMF) vii 62 64 65 69 69 70 70 71 73 73 75 75 80 80 85 87 89 89 92 95 96 97 102 104 105 107 108 113 114 viii CONTENTS 5.5.2 Iteratively Weighted Least Squares (IWLS) 5.5.3 NNMF: Gaussian Case (Approximate Noise) 5.5.4 Factored NNMF: Gaussian Case (Approximate Data) 5.5.5 Alternating Least Squares (ALS) 5.5.6 Performance Comparisons 5.6 Conclusions Acknowledgements References Hyperspectral Image Data Conditioning and Regression Analysis James E Burger and Paul L M Geladi 6.1 Introduction 6.2 Terminology 6.3 Multivariate Image Regression 6.3.1 Regression Diagnostics 6.3.2 Differences between Normal Calibration and Image Calibration 6.4 Data Conditioning 6.4.1 Reflectance Transformation and Standardization 6.4.2 Spectral Transformations 6.4.3 Data Clean-up 6.4.4 Data Conditioning Summary 6.5 PLS Regression Optimization 6.5.1 Data Subset Selection 6.5.2 Pseudorank Determination 6.6 Regression Examples 6.6.1 Artificial Ternary Mixture 6.6.2 Commercial Cheese Samples 6.6.3 Wheat Straw Wax 6.7 Conclusions Acknowledgements References Principles of Image Cross-validation (ICV): Representative Segmentation of Image Data Structures Kim H Esbensen and Thorbjørn T Lied 7.1 Introduction 7.2 Validation Issues 117 118 119 120 121 124 125 125 127 127 128 128 130 132 132 133 135 137 138 138 138 139 140 142 146 149 150 152 152 155 155 156 CONTENTS 7.2.1 2-way 7.2.2 MIA/MIR 7.3 Case Studies 7.3.1 Case 1: Full Y-image 7.3.2 Case 2: Critical Segmentation Issues 7.3.3 Case 3: Y-composite 7.3.4 Case 4: Image Data Structure Sampling 7.4 Discussion and Conclusions 7.5 Reflections on 2-way Cross-validation References Detection, Classification, and Quantification in Hyperspectral Images Using Classical Least Squares Models Neal B Gallagher 8.1 Introduction 8.2 CLS Models 8.2.1 CLS 8.2.2 ELS 8.2.3 GLS 8.3 Detection, Classification, and Quantification 8.3.1 Detection 8.3.2 Classification 8.3.3 Quantification 8.4 Conclusions Acknowledgements References ix 156 158 160 163 165 168 173 177 178 180 181 181 182 183 187 189 192 194 196 197 200 200 201 Calibration Standards and Image Calibration Paul L M Geladi 203 9.1 Introduction 9.2 The Need for Calibration in General 9.3 The Need for Image Calibration 9.4 Resolution in Hyperspectral Images 9.5 Spectroscopic Definitions 9.6 Calibration Standards 9.7 Calibration in Hyperspectral Images 9.8 Conclusions References 203 203 204 205 207 209 213 219 219 x CONTENTS 10 Multivariate Movies and their Applications in Pharmaceutical and Polymer Dissolution Studies Jaap van der Weerd and Sergei G Kazarian 10.1 Introduction 10.1.1 Introducing the Time Axis 10.1.2 Data Structure and Reduction 10.1.3 Compression of Spectra 10.1.4 Space Dimensions 10.1.5 Time Dimension 10.1.6 Simultaneous Compression of all Variables 10.2 Applications: Solvent Diffusion and Pharmaceutical Studies 10.2.1 Solvent Diffusion in Polymers 10.2.2 Optical and NMR Studies 10.2.3 Line Imaging 10.2.4 Global MIR Imaging Studies of Solvent Intake 10.3 Drug Release 10.3.1 ATR-FTIR Imaging 10.4 Conclusions Acknowledgement References 11 Multivariate Image Analysis of Magnetic Resonance Images: Component Resolution with the Direct Exponential Curve Resolution Algorithm (DECRA) Brian Antalek, Willem Windig and Joseph P Hornak 11.1 11.2 11.3 11.4 11.5 11.6 11.7 Introduction DECRA Approach DECRA Algorithm H Relaxation T1 Transformation Imaging Methods Phantom Images 11.7.1 T2 Series 11.7.2 T1 Series 11.8 Brain Images 11.8.1 T2 Series 11.8.2 T1 Series 11.9 Regression Analysis 11.10 Conclusions References 221 221 222 223 224 227 231 235 237 238 242 245 246 249 251 254 255 255 261 261 264 269 270 271 271 273 273 277 278 278 281 282 285 285 364 Clustering 230–1, 232 fuzzy c-means 46–9 cluster size insensitive 49–51, 64 spatially guided 52–4 and near infrared chemical imaging 352 Clutter 186 extended least squares model 187–8 generalized least squares model 189–92 Color imaging 2–3, 9, 10 class coloring 62–3, 65 Color-slice contouring 24 Component analysis, see under Count data analysis Compound feed, see Animal feed Compression all variables simultaneously 235–7 space dimensions 227–31 spectra 224–7 time dimensions 231–5 Conditioning, see Data conditioning Contaminant detection cereals 301–2 compound feed 302–4 fruit 298–301 meat products 297–8 NIR chemical imaging 348–9 Correlogram 234 Count data analysis 89–123 background to spectral analysis 90–5 component analysis 95–6 maximum likelihood approaches 113–24 method performance comparisons 121–5 orthogonal matrix factorization 96–113 example datasets and simulations 92–5 related spectroscopic techniques 90 Cross-validation defined 157 model test set 131 multivariate analysis/regression 158–60 two-way chemometrics 156–8, 178–80 see also Representative segmentation D-metric 139–40, 146, 148 Data conditioning 132–8 data clean-up 137–8 reflectance transformations/standardization 133–5 spectral transformations 135–7 standard reference materials 133, 134 INDEX standardization defined 128 transformation defined 128 Data pre-processing 127, 135–6, 320–2 Data quality 127–8, 150 DECRA (NMR imaging) 261–88 algorithm 269–70 approach 264–9 background to MRI 261–4 brain scan image 373 T1 series 281–2 T2 series 278–80 H relaxation 270 image acquisitions 271–3 phantom test image 271–2 T1 series 277–8 T2 series 273–6 regression analysis 282–5 transformation T1 271 Diffusion, see Solvent diffusion Digital images 1–3 Direct exponential curve resolution algorithm, see DECRA (NMR imaging) Drug studies chemical heterogeneity/homogeneity 336–7, 358–9 dissolution 241–5, 249–54 distribution of time-release beads 357 Eigenanalysis 97–8 spectral variance and noise 109–13 Energy dispersive X-ray spectrometer 92 Extended least squares model 181, 187–8 Factor analysis 96–7 FCM, see Fuzzy c-means techniques Feed mixtures 296 contamination in 302–4 Feedback multivariate model selection (FEMOS) 56–7 multispectral image of pine and spruce 58–60 as unsupervised clustering technique 60–2 Fick’s laws 238–40 Field of view 134, 348 Focal plane scanning Food analysis techniques 291–304 animal feed mixtures 296, 302–4 cheese 146–9 fruit and vegetables 50, 54, 291–4, 298–301 INDEX kernels/cereals 294–5, 301–2 meat products 51, 297–8 see also Quality inspection (agricultural products) Fruit and vegetable analysis 291–4 defects and contaminants 50, 54, 298–301 FTIR studies drug release 249–54 polymer dissolution 246–9 reflectance data of aspirin/sugar mixture 71, 80 transmission data of polymer laminate 70–1, 75–9 Fuzzy c-means techniques 46–8, 64 cluster size insensitive 49–51, 64 conditional 48–9 spatially guided 51–4, 64 Generalized least squares model 182, 189 Generalized rank annihilation method 263, 265, 268 Geometric standards 209–10, 211 Grayscale 211 Grayvalues 2–3, 206 Guided random sampling, see under Representative segmentation Histograms colour-slice contouring 24 of pixel intensities 355–6 Hotelling’s T statistic 184, 185, 189 Hyperspectral image defined generation of 5–9 Image acquisition 55 camera configurations 6–9, 349 InGaAs camera with LCTF filter 213–19 magnetic resonance imaging 271–3 Image cross-validation, see Cross-validation Image regression, see Multivariate image regression Image subclasses 27–8 InGaAs camera 213–19 Intensity calibration 204 grayvalues 2–3, 206 Interference 181, 186 extended least squares model 187–8 365 generalized least squares model 181, 189–92 Internal reference standards 127 Inverse least squares 181, 226 Iteratively weighted least squares 117–18 Joint normal distribution 26–7 Kernels/cereals analysis 294–5, 301–2 Kubelka–Munk transform 136, 183 Least squares alternating 120–1 extended 181, 187–8 generalized 182, 189–92 iteratively weighted 117–19 partial 129–30, 138–40, 181 see also Classical least squares models Leave-one-out-cross-validation 179 Line scanning 7–8 Raman spectroscopy 245–6 Liquid crystal tunable filter 213–14 Local modeling 27–8 Magnetic resonance imaging, see DECRA (NMR imaging) Maltese cross segmentation 163, 164, 165–8, 171, 173, 177–8 Matlab 95 function 103–4 Maximum likelihood analysis 113–21 maximum likelihood PCA 104–5 method performance comparisons 121–4 Meat products 51 defects and contaminants 297–8 Mixing class concept 33–8 Montmorency Forest data set 18–21, 160–77 multivariate image analysis 21–4 representative segmentation 160–77 MRI, see DECRA (NMR imaging) Multiplicative scatter correction 128, 136, 146 Multiplicative signal correction 137 Multivariate curve resolution 186 Multivariate image defined 3–4 relationship to hyperspectral image 5, 181 366 Multivariate image analysis, principles of 17–42 basics illustrated 9–14 brushing 24–6 colour-slice contouring 24 local modeling and image subclasses 27–8 mixing class concept 33–8 primary classes 32–3 scene space iteration with score space 32–3 sampling in score space 39–40 score space and score cross-plot 21–4 synopsis 18, 40–1 topographic map analogy 28–31 validation issues 158–9 see also Count data analysis Multivariate image regression, see Regression analysis Multivariate movies 207, 221–60 applications (diffusion/drug studies) 237–55 background discussion 237–42 drug release (FTIR/ARC) 249–54 line imaging 245–6 MIR solvent intake studies 246–9 optical and NMR studies 242–5 background to 121–2 data structure and compression 223–4 all variables simultaneously 235–7 space dimensions 227–31 spectra 224–7 time dimension 231–5 Near infrared chemical imaging 335–61 acquisition schemes/ measurement modes 341–2 background and scope 335–8, 343–4 chemical contrast and classification procedures 350–4 data analysis spatial and spectral unmixing 346–8 underlying data structure 344–6 histograms of pixel intensities 355–6 instrumentation and data collection 338–40 sample heterogeneity and type 340–1 performance/behavior correlations 359–60 spatial dimensions 348–50 spectral data INDEX homogeneity measurement 358–9 statistics and spatial relationships 357–8 NIST wavelength standards 213 NMR multivariate movies 242–5 see also DECRA (NMR imaging) Noise Poisson 92, 95, 99–102, 107, 120 pre-normalization for PET images 321–2 see also Interference Non-negative matrix factorization 118–20 Nuclear magnetic resonance biological imaging, see DECRA (NMR imaging) multivariate movies 242–5 Nuclear medicine, see Positron emission tomography Orthogonal matrix factorization Outliers 137 96–113 Palmitic/stearic acid mixture 71–3, 80–5 Parafac model 236 Partial least squares 129–30, 181 data subset selection 138–9 and NIR chemical imaging 353–4 pseudorank determination 139–40 Particle tracking 231, 233 Pepper images 54 PET, see Positron emission tomography Pharmaceuticals, see Drug studies Piecewise multiplicative scatter correction 137, 146 Pixels dead/bad 137, 184 defined histograms of intensities 355–6 size 205–6 Point scanning 6–7, 349 Poisson distribution 91–2 Poisson noise 92, 95, 99–102, 107, 120 Poisson non-negative matrix factorization 114–17, 120 Polyethylene/aspirin mixture 193–200 Polymers laminate 70–1, 75–9, 75–80 solvent diffusion in 237–41, 245–9 Positron emission tomography 313–34 background to imaging 313–15 basic principles 315–17 INDEX 367 data and image analysis methods 318–19 data pre-processing noise-prenormalization 321–2 scaling and mean-centring 320–1 dynamic and static scanning modes 317 masked volume wise PCA 326, 331 masking procedure 327–8 sinogram domain application 328–30 PCA basics 319–20, 322–3 slice wise/sinogram wise PCA 323, 331 contrast/kinetic normalization 325–6 noise pre-normalization 323–4 outlining reference region 324–5 volume wise PCA 326 Potato images 47, 50 Poultry 298 Pre-processing, see Data pre-processing Prediction performance 131, 156–7 Principal component analysis 10, 97, 226–7, 319–20 application to PET studies 314, 320–32 arbitrary factor models 102–4 data scaling and mean-centring 20–1 definition 319–20 and eigenanalysis 97–8 history of 319 maximum likelihood PCA 104–5 and near infrared chemical imaging 351–2 principal factor analysis 107–8 selecting the number of components 108–13 and singular value decomposition 97–8 training set selection 55–6 weighted PCA 105–7 see also Multivariate image analysis, principles of Principal factor analysis 107–8 Proton spin relaxation 270 Pseudoabsorbance 209 Pseudocolors 4, 10, 11, 62–3 Pseudorank 139 Pulsed-gradient spin echo data 263, 264–9 unsupervised (fuzzy c-means) 46–54 techniques reviewed 44–5 cereals 301–2 compound feed 302–4 fruit 298–301 meat products 297–8 see also Food analysis techniques Raman spectroscopy aspirin/polyethylene mixture 193–200 line imaging 245–6 Reflectance defined 208 specular on chemical images 342 standards 210–13 transformations 133–5 Regression analysis calibration 128–31 image versus normal 132 and cross-validation 158–60 diagnostics 130–2 example datasets 140–2, 151–2 cheese samples 146–9 citric acid/salicylic acid/sugar mixture 142–6 and interference 181 MRI images 282–5 and NIR chemical imaging 353 partial least squares 129–30 data subset selection 138–9 pseudorank determination 139 wheat straw wax 149–50 Remote sensing 222, 289, 290, 304–7 Representative segmentation ‘blind’ approaches 155, 159, 165, 179 guided random sampling (case study) 160–3 critical segmentation issues 165–8 data structure and sampling 173–7 full Y-image 163–5 Y-composite 168–73 Root mean square error of calibration 131 of prediction 131, 156 Quality inspection (agricultural products) 43–67 class coloring and pseudo color 62–3, 65 consumer perceptions of quality 43–4 image clustering/classification supervised (FEMOS) 55–7 Satellite systems 222, 289, 290, 304–7 Scanning electron microscopy 92–5 Scene space 17, 18, 31 defined 10 sampling in score space 39–40 and score space iteration 32 Score plot 10–11, 12, 13, 18, 21–4, 30 368 Score space 10, 18, 21–4, 30, 39–40 and scene space iteration 32 scene space sampling in 39–40 Secondary ion mass spectrometry palmitic/stearic acid mixture 71–3, 80–5 TOF-SIMS of plated copper grid 92–5 Segmentation, see Representative segmentation Self-modeling image analysis (SIMPLISMA) 69–87 FTIR data reflectance (aspirin/sugar) 71, 80 transmission (polymer laminate) 70–1, 75–9 SIMS of palmitic/stearic acid mixture 71–3, 80–5 theoretical background 70, 73–5 Singular value decomposition 97–8 Size of image 205, 206 Solvent diffusion drug release studies 241–5, 249–54 in polymers 237–41, 245–9 Spatial heterogeneity 340 Spatially guided FCM 51–4, 64 Spectral resolution 206–7 Spectral transformations 135–7 Spectralon 133–5, 210, 212 Spectrophotometer 5–6 Spectroscopic analysis 90, 183 NIR spectral data 305–9 see also Count data analysis Spin-lattice/spin-spin relaxation times 261–4 and DECRA 269, 270–1 Standard normal variate transform 137 Standardization 128 INDEX reference materials 133, 134 see also Calibration Stearic/palmitic acid mixture 71–3, 80–5 Sugar/aspirin mixture 71, 80 Supervised classification 17, 54–6, 353–4 Templar cross, see Maltese cross segmentation Temporal images, see Multivariate movies Test set selection 55–6 validation 157, 180 Time resolved images, see Multivariate movies Time series analysis 231–5 Tissue samples 261–4 TOF-SIMS of plated copper grid 92–5 Training set selection 55–6 Transformation 128, 133–7 Transmittance defined 207–8 Tucker3 model 236 Two-way cross-validation 156–8, 178–80 Unsupervised classification 17, 46–53, 350–2 User-delineated segmentation 155 Validation 131, 156–8 Vegetables, see Fruit and vegetable analysis Voxel 2, 206 Wavelength resolution 206–7 Weighted PCA 105–7 Wheat straw wax 149–50 Wood images 59 Plate Plate Plate Plate (a) (b) Plate Techniques and Applications of Hyperspectral Image Analysis Edited by H F Grahn and P Geladi © 2007 John Wiley & Sons, Ltd ISBN: 978-0-470-01086-0 (a) (b) (c) Plate Plate Plate Plate (a) (b) Plate 10 Plate 11 Plate 13 Plate 12 (a) (b) (d) Plate 14 Plate 16 Plate 15 (c) (e) Plate 17 Plate 18 (a) (b) Line Line Plate 19 Plate 20 Plate 21 Plate 22 (a) Plate 23 (b) (a) (b) Plate 24 Plate 25 Plate 27 Plate 26 Plate 28 (a) (c) Plate 29 (b) (d) Plate 30 LD Polyethylene HD Polyethylene 10 12 14 16 18 20 10 15 20 Aspirin 25 30 Plate 31 HE staining clusters A clusters clusters D Wards algorithm D-values clusters B C 11 clusters E F second derivatives (vector-normalized) 2 4 5 6 7 8 10 G Plate 32 10 11 H 11 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 wavenumber (cm–1) (a) Position / µm 60 50 40 30 20 10 20 30 Time / 40 50 10 20 30 Time / 40 50 10 20 30 Time / 40 50 (b) Position / µm 60 50 40 30 20 (c) Position / µm 60 50 40 30 20 Plate 33 Plate 35 Plate 34 .. .Techniques and Applications of Hyperspectral Image Analysis Techniques and Applications of Hyperspectral Image Analysis Edited by H F Grahn and P Geladi © 2007 John Wiley... but is often and preferably quantitative This Techniques and Applications of Hyperspectral Image Analysis Edited by H F Grahn and P Geladi © 2007 John Wiley & Sons, Ltd ISBN: 97 8-0 -4 7 0-0 108 6-0 MULTIVARIATE... the concepts of digital image, multivariate image and hyperspectral image and gives an overview of some of the image generation techniques for producing multivariate and hyperspectral images The