计算机与机器视觉理论、算法与实践 英文版 第4版2025|PDF|Epub|mobi|kindle电子书版本百度云盘下载

计算机与机器视觉理论、算法与实践 英文版 第4版PDF格式电子书版下载

注意：本站所有压缩包均有解压码： 点击下载压缩包解压工具

CHAPTER 1 Vision,the Challenge1

1.1 Introduction―Man and His Senses1

1.2 The Nature of Vision2

1.2.1 The Process of Recognition2

1.2.2 Tackling the Recognition Problem4

1.2.3 Object Location6

1.2.4 Scene Analysis8

1.2.5 Vision as Inverse Graphics9

1.3 From Automated Visual Inspection to Surveillance10

1.4 What This Book is About12

1.5 The Following Chapters13

1.6 Bibliographical Notes14

PART 1 LOW-LEVEL VISION15

CHAPTER 2 Images and Imaging Operations17

2.1 Introduction18

2.1.1 Gray Scale Versus Color19

2.2 Image Processing Operations23

2.2.1 Some Basic Operations on Grayscale Images24

2.2.2 Basic Operations on Binary Images28

2.3 Convolutions and Point Spread Functions32

2.4 Sequential Versus Parallel Operations34

2.5 Concluding Remarks36

2.6 Bibliographical and Historical Notes36

2.7 Problems36

CHAPTER 3 Basic Image Filtering Operations38

3.1 Introduction38

3.2 Noise Suppression by Gaussian Smoothing40

3.3 Median Filters43

3.4 Mode Filters45

3.5 Rank Order Filters52

3.6 Reducing Computational Load54

3.7 Sharp-Unsharp Masking55

3.8 Shifts Introduced by Median Filters56

3.8.1 Continuum Model of Median Shifts57

3.8.2 Generalization to Grayscale Images59

3.8.3 Problems with Statistics60

3.9 Discrete Model of Median Shifts62

3.10 Shifts Introduced by Mode Filters65

3.11 Shifts Introduced by Mean and Gaussian Filters67

3.12 Shifts Introduced by Rank Order Filters68

3.12.1 Shifts in Rectangular Neighborhoods69

3.13 The Role of Filters in Industrial Applications of Vision74

3.14 Color in Image Filtering74

3.15 Concluding Remarks76

3.16 Bibliographical and Historical Notes77

3.16.1 More Recent Developments78

3.17 Problems79

CHAPTER 4 Thresholding Techniques82

4.1 Introduction83

4.2 Region-Growing Methods83

4.3 Thresholding84

4.3.1 Finding a Suitable Threshold85

4.3.2 Tackling the Problem of Bias in Threshold Selection86

4.3.3 Summary88

4.4 Adaptive Thresholding88

4.4.1 The Chow and Kaneko Approach91

4.4.2 Local Thresholding Methods92

4.5 More Thoroughgoing Approaches to Threshold Selection93

4.5.1 Variance-Based Thresholding95

4.5.2 Entropy-Based Thresholding96

4.5.3 Maximum Likelihood Thresholding97

4.6 The Global Valley Approach to Thresholding98

4.7 Practical Results Obtained Using the Global ValleyMethod101

4.8 Histogram Concavity Analysis106

4.9 Concluding Remarks107

4.10 Bibliographical and Historical Notes108

4.10.1 More Recent Developments109

4.11 Problems110

CHAPTER 5 Edge Detection111

5.1 Introduction112

5.2 Basic Theory of Edge Detection113

5.3 The Template Matching Approach115

5.4 Theory of 3×3 Template Operators116

5.5 The Design of Differential Gradient Operators117

5.6 The Concept of a Circular Operator118

5.7 Detailed Implementation of Circular Operators120

5.8 The Systematic Design of Differential Edge Operators122

5.9 Problems with the Above Approach―Some Alternative Schemes123

5.10 Hysteresis Thresholding126

5.11 The Canny Operator128

5.12 The Laplacian Operator131

5.13 Active Contours134

5.14 Practical Results Obtained Using Active Contours137

5.15 The Level Set Approach to Object Segmentation140

5.16 The Graph Cut Approach to Object Segmentation141

5.17 Concluding Remarks145

5.18 Bibliographical and Historical Notes146

5.18.1 More Recent Developments147

5.19 Problems148

CHAPTER 6 Corner and Interest Point Detection149

6.1 Introduction150

6.2 Template Matching150

6.3 Second-Order Derivative Schemes151

6.4 A Median Filter-Based Corner Detector153

6.4.1 Analyzing the Operation of the Median Detector154

6.4.2 Practical Results156

6.5 The Harris Interest Point Operator158

6.5.1 Corner Signals and Shifts for Various Geometric Configurations161

6.5.2 Performance with Crossing Points and Junctions162

6.5.3 Different Forms of the Harris Operator165

6.6 Corner Orientation166

6.7 Local Invariant Feature Detectors and Descriptors168

6.7.1 Harris Scale and Affine-Invariant Detectors and Descriptors171

6.7.2 Hessian Scale and Affine-Invariant Detectors and Descriptors173

6.7.3 The SIFT Operator173

6.7.4 The SURF Operator174

6.7.5 Maximally Stable Extremal Regions176

6.7.6 Comparison of the Various Invariant Feature Detectors177

6.8 Concluding Remarks180

6.9 Bibliographical and Historical Notes181

6.9.1 More Recent Developments184

6.10 Problems184

CHAPTER 7 Mathematical Morphology185

7.1 Introduction185

7.2 Dilation and Erosion in Binary Images186

7.2.1 Dilation and Erosion186

7.2.2 Cancellation Effects186

7.2.3 Modified Dilation and Erosion Operators187

7.3 Mathematical Morphology187

7.3.1 Generalized Morphological Dilation187

7.3.2 Generalized Morphological Erosion188

7.3.3 Duality Between Dilation and Erosion189

7.3.4 Properties of Dilation and Erosion Operators190

7.3.5 Closing and Opening193

7.3.6 Summary of Basic Morphological Operations195

7.4 Grayscale Processing197

7.4.1 Morphological Edge Enhancement198

7.4.2 Further Remarks on the Generalization to Grayscale Processing199

7.5 Effect of Noise on Morphological Grouping Operations201

7.5.1 Detailed Analysis203

7.5.2 Discussion205

7.6 Concluding Remarks205

7.7 Bibliographical and Historical Notes206

7.7.1 More Recent Developments207

7.8 Problem208

CHAPTER 8 Texture209

8.1 Introduction209

8.2 Some Basic Approaches to Texture Analysis213

8.3 Graylevel Co-occurrence Matrices213

8.4 Laws'Texture Energy Approach217

8.5 Ade's Eigenfilter Approach220

8.6 Appraisal of the Laws and Ade Approaches221

8.7 Concluding Remarks223

8.8 Bibliographical and Historical Notes223

8.8.1 More Recent Developments224

PART 2 INTERMEDIATE-LEVEL VISION227

CHAPTER 9 Binary Shape Analysis229

9.1 Introduction230

9.2 Connectedness in Binary Images230

9.3 Object Labeling and Counting231

9.3.1 Solving the Labeling Problem in a More Complex Case235

9.4 Size Filtering238

9.5 Distance Functions and Their Uses240

9.5.1 Local Maxima and Data Compression243

9.6 Skeletons and Thinning244

9.6.1 Crossing Number247

9.6.2 Parallel and Sequential Implementations of Thinning248

9.6.3 Guided Thinning251

9.6.4 A Comment on the Nature of the Skeleton251

9.6.5 Skeleton Node Analysis251

9.6.6 Application of Skeletons for Shape Recognition253

9.7 Other Measures for Shape Recognition254

9.8 Boundary Tracking Procedures257

9.9 Concluding Remarks257

9.10 Bibliographical and Historical Notes259

9.10.1 More Recent Developments260

9.11 Problems261

CHAPTER 10 Boundary Pattern Analysis266

10.1 Introduction266

10.2 Boundary Tracking Procedures269

10.3 Centroidal Profiles269

10.4 Problems with the Centroidal Profile Approach270

10.4.1 Some Solutions271

10.5 The(s,ψ)Plot274

10.6 Tackling the Problems of Occlusion276

10.7 Accuracy of Boundary Length Measures279

10.8 Concluding Remarks280

10.9 Bibliographical and Historical Notes281

10.9.1 More Recent Developments282

10.10 Problems282

CHAPTER 11 Line Detection284

11.1 Introduction284

11.2 Application of the Hough Transform to Line Detection285

11.3 The Foot-of-Normal Method288

11.3.1 Application of the Foot-of-Normal Method290

11.4 Longitudinal Line Localization290

11.5 Final Line Fitting292

11.6 Using RANSAC for Straight Line Detection293

11.7 Location of Laparoscopic Tools297

11.8 Concluding Remarks299

11.9 Bibliographical and Historical Notes300

11.9.1 More Recent Developments301

11.10 Problems301

CHAPTER 12 Circle and Ellipse Detection303

12.1 Introduction304

12.2 Hough-Based Schemes for Circular Object Detection305

12.3 The Problem of Unknown Circle Radius308

12.3.1 Some Practical Results310

12.4 The Problem ofAccurate Center Location311

12.4.1 A Solution Requiring Minimal Computation313

12.5 Overcoming the Speed Problem314

12.5.1 More Detailed Estimates of Speed314

12.5.2 Robustness315

12.5.3 Practical Results316

12.5.4 Summary317

12.6 Ellipse Detection320

12.6.1 The Diameter Bisection Method320

12.6.2 The Chord-Tangent Method322

12.6.3 Finding the Remaining Ellipse Parameters323

12.7 Human Iris Location325

12.8 Hole Detection327

12.9 Concluding Remarks327

12.10 Bibliographical and Historical Notes328

12.10.1 More Recent Developments330

12.11 Problems331

CHAPTER 13 The Hough Transform and Its Nature333

13.1 Introduction333

13.2 The Generalized Hough Transform334

13.3 Setting Up the Generalized Hough Transform—Some Relevant Questions336

13.4 Spatial Matched Filtering in Images336

13.5 From Spatial Matched Filters to Generalized Hough Transforms337

13.6 Gradient Weighting Versus Uniform Weighting339

13.6.1 Calculation of Sensitivity and Computational Load339

13.7 Summary342

13.8 Use of the GHT for Ellipse Detection343

13.8.1 Practical Details347

13.9 Comparing the Various Methods349

13.10 Fast Implementations of the Hough Transform350

13.11 The Approach of Gerig and Klein352

13.12 Concluding Remarks353

13.13 Bibliographical and Historical Notes354

13.13.1 More Recent Developments356

13.14 Problems357

CHAPTER 14 Pattern Matching Techniques358

14.1 Introduction359

14.2 A Graph-Theoretic Approach to Object Location359

14.2.1 A Practical Example―Locating Cream Biscuits363

14.3 Possibilities for Saving Computation366

14.4 Using the Generalized Hough Transform for Feature Collation369

14.4.1 Computational Load370

14.5 Generalizing the Maximal Clique and Other Approaches371

14.6 Relational Descriptors373

14.7 Search376

14.8 Concluding Remarks377

14.9 Bibliographical and Historical Notes378

14.9.1 More Recent Developments380

14.10 Problems381

PART 3 3-D VISION AND MOTION387

CHAPTER 15 The Three-Dimensional World389

15.1 Introduction389

15.2 3-D Vision—the Variety of Methods390

15.3 Projection Schemes for Three-Dimensional Vision392

15.3.1 Binocular Images393

15.3.2 The Correspondence Problem396

15.4 Shape from Shading398

15.5 Photometric Stereo402

15.6 The Assumption of Surface Smoothness405

15.7 Shape from Texture407

15.8 Use of Structured Lighting408

15.9 Three-Dimensional Object Recognition Schemes410

15.10 Horaud's Junction Orientation Technique411

15.11 An Important Paradigm―Location of Industrial Parts415

15.12 Concluding Remarks417

15.13 Bibliographical and Historical Notes419

15.13.1 More Recent Developments420

15.14 Problems421

CHAPTER 16 Tackling the Perspective n-point Problem424

16.1 Introduction424

16.2 The Phenomenon of Perspective Inversion425

16.3 Ambiguity of Pose under Weak Perspective Projection427

16.4 Obtaining Unique Solutions to the Pose Problem430

16.4.1 Solution of the Three-Point Problem433

16.4.2 Using Symmetric Trapezia for Estimating Pose434

16.5 Concluding Remarks434

16.6 Bibliographical and Historical Notes436

16.6.1 More Recent Developments437

16.7 Problems438

CHAPTER 17 Invariants and Perspective439

17.1 Introduction440

17.2 Cross-ratios:the“Ratio of Ratios”Concept441

17.3 Invariants for Noncollinear Points445

17.3.1 Further Remarks About the Five-Point Configuration447

17.4 Invariants for Points on Conics449

17.5 Differential and Semi-difierential Invariants452

17.6 Symmetric Cross-ratio Functions454

17.7 Vanishing Point Detection456

17.8 More on Vanishing Points458

17.9 Apparent Centers of Circles and Ellipses460

17.10 The Route to Face Recognition462

17.10.1 The Face as Part of a 3-D Object464

17.11 Perspective Effects in Art and Photography466

17.12 Concluding Remarks472

17.13 Bibliographical and Historical Notes474

17.13.1 More Recent Developments475

17.14 Problems475

CHAPTER 18 Image Transformations and Camera Calibration478

18.1 Introduction479

18.2 Image Transformations479

18.3 Camera Calibration483

18.4 Intfinsic and Extrinsic Parameters486

18.5 Correcting for Radial Distortions488

18.6 Multiple View Vision490

18.7 Generalized Epipolar Geometry491

18.8 The Essential Matrix492

18.9 The Fundamental Matrix495

18.10 Properties of the Essential and Fundamental Matrices496

18.11 Estimating the Fundamental Matrix497

18.12 An Update on the Eight-Point Algorithm497

18.13 Image Rectification498

18.14 3-D Reconstruction499

18.15 Concluding Remarks501

18.16 Bibliographical and Historical Notes502

18.16.1 More Recent Developments503

18.17 Problems504

CHAPTER 19 Motion505

19.1 Introduction505

19.2 Optical Flow506

19.3 Interpretation of Optical Flow Fields509

19.4 Using Focus of Expansion to Avoid Collision511

19.5 Time-to-Adjacency Analysis513

19.6 Basic Difficulties with the Optical Flow Model514

19.7 Stereo from Motion515

19.8 The Kalman Filter517

19.9 Wide Baseline Matching519

19.10 Concluding Remarks521

19.11 Bibliographical and Historical Notes522

19.12 Problem522

PART 4 TOWARD REAL-TIME PATTERN RECOGN ITION SYSTEMS523

CHAPTER 20 Automated Visual Inspection525

20.1 Introduction525

20.2 The Process of Inspection527

20.3 The Types of Object to be Inspected527

20.3.1 Food Products528

20.3.2 Precision Components528

20.3.3 Differing Requirements for Size Measurement529

20.3.4 Three-Dimensional Objects530

20.3.5 Other Products and Materials for Inspection530

20.4 Summary:The Main Categories of Inspection530

20.5 Shape Deviations Relative to a Standard Template532

20.6 Inspection of Circular Products533

20.7 Inspection of Printed Circuits537

20.8 Steel Strip and Wood Inspection538

20.9 Inspection of Products with High Levels of Variability539

20.10 X-Ray Inspection542

20.10.1 The Dual-Energy Approach to X-Ray Inspection546

20.11 The Importance of Color in Inspection546

20.12 Bringing Inspection to the Factory548

20.13 Concluding Remarks549

20.14 Bibliographical and Historical Notes550

20.14.1 More Recent Developments552

CHAPTER 21 Inspection of Cereal Grains553

21.1 Introduction553

21.2 Case Study:Location of Dark Contaminants in Cereals554

21.2.1 Application of Morphological and Nonlinear Filters to Locate Rodent Droppings555

21.2.2 Problems with Closing558

21.2.3 Ergot Detection Using the Global Valley Method558

21.3 Case Study:Location of Insects560

21.3.1 The Vectorial Strategy for Linear Feature Detection560

21.3.2 Designing Linear Feature Detection Masks for Larger Windows563

21.3.3 Application to Cereal Inspection564

21.3.4 Experimental Resuits564

21.4 Case Study:High-Speed Grain Location566

21.4.1 Extending an Earlier Sampling Approach566

21.4.2 Application to Grain Inspection567

21.4.3 Summary571

21.5 Optimizing the Output for Sets of Directional Template Masks572

21.5.1 Application of the Formulae573

21.5.2 Discussion574

21.6 Concluding Remarks575

21.7 Bibliographical and Historical Notes575

21.7.1 More Recent Developments576

CHAPTER 22 Surveillance578

22.1 Introduction579

22.2 Surveillance—The Basic Geometry580

22.3 Foreground―Background Separation584

22.3.1 Background Modeling585

22.3.2 Practical Examples of Background Modeling591

22.3.3 Direct Detection of the Foreground593

22.4 Particle Filters594

22.5 Use of Color Histograms for Tracking600

22.6 Implementation of Particle Filters604

22.7 Chamfer Matching,Tracking,and Occlusion607

22.8 Combining Views from Multiple Cameras609

22.8.1 The Case of Nonoverlapping Fields of View613

22.9 Applications to the Monitoring of Traffic Flow614

22.9.1 The System of Bascle et al614

22.9.2 The System of Koller et al616

22.10 License Plate Location619

22.11 Occlusion Classification for Tracking621

22.12 Distinguishing Pedestrians by Their Gait623

22.13 Human Gait Analysis627

22.14 Model-Based Tracking of Animals629

22.15 Concluding Remarks631

22.16 Bibliographical and Historical Notes632

22.16.1 More Recent Developments634

22.17 Problem635

CHAPTER 23 In-Vehicle Vision Systems636

23.1 Introduction637

23.2 Locating the Roadway638

23.3 Location of Road Markings640

23.4 Location of Road Signs641

23.5 Location of Vehicles645

23.6 Information Obtained by Viewing License Plates and Other Structural Features647

23.7 Locating Pedestrians651

23.8 Guidance and Egomotion653

23.8.1 A Simple Path Planning Algorithm656

23.9 Vehicle Guidance in Agriculture656

23.9.1 3-D Aspects of the Task660

23.9.2 Real-Time Implementation661

23.10 Concluding Remarks662

23.11 More Detailed Developments and Bibliographies Relating to Advanced Driver Assistance Systems663

23.11.1 Developments in Vehicle Detection664

23.11.2 Developments in Pedestrian Detection666

23.11.3 Developments in Road and Lane Detection668

23.11.4 Developments in Road Sign Detection669

23.11.5 Developments in Path Planning,Navigation, and Egomotion671

23.12 Problem671

CHAPTER 24 Statistical Pattern Recognition672

24.1 Introduction673

24.2 The Nearest Neighbor Algorithm674

24.3 Bayes'Decision Theory676

24.3.1 The Naive Bayes'Classifier678

24.4 Relation of the Nearest Neighbor and Bayes' Approaches679

24.4.1 Mathematical Statement of the Problem679

24.4.2 The Importance of the Nearest Neighbor Classifier681

24.5 The Optimum Number of Features681

24.6 Cost Functions and Error-Reject Tradeoff682

24.7 The Receiver Operating Characteristic684

24.7.1 On the Variety of Performance Measures Relating to Error Rates686

24.8 Multiple Classifiers688

24.9 Cluster Analysis691

24.9.1 Supervised and Unsupervised Learning691

24.9.2 Clustering Procedures692

24.10 Principal Components Analysis695

24.11 The Relevance of Probability in Image Analysis699

24.12 Another Look at Statistical Pattern Recognition:The Support Vector Machine700

24.13 Artificial Neural Networks701

24.14 The Back-Propagation Algorithm705

24.15 MLP Architectures708

24.16 Overfitting to the Training Data709

24.17 Concluding Remarks712

24.18 Bibliographical and Historical Notes713

24.18.1 More Recent Developments715

24.19 Problems717

CHAPTER 25 Image Acquisition718

25.1 Introduction718

25.2 Illumination Schemes719

25.2.1 Eliminating Shadows721

25.2.2 Principles for Producing Regions of Uniform Illumination724

25.2.3 Case of Two Infinite Parallel Strip Lights726

25.2.4 Overview of the Uniform Illumination Scenario729

25.2.5 Use of Line-Scan Cameras730

25.2.6 Light Emitting Diode(LED)Sources731

25.3 Cameras and Digitization732

25.3.1 Digitization734

25.4 The Sampling Theorem735

25.5 Hyperspectral Imaging738

25.6 Concluding Remarks739

25.7 Bibliographical and Historical Notes740

25.7.1 More Recent Developments741

CHAPTER 26 Real-Time Hardware and Systems Design Considerations742

26.1 Introduction743

26.2 Parallel Processing744

26.3 SIMD Systems745

26.4 The Gain in Speed Attainable with N Processors747

26.5 Flynn's Classification748

26.6 Optimal Implementation of Image Analysis Algorithms750

26.6.1 Hardware Specification and Design751

26.6.2 Basic Ideas on Optimal Hardware Implementation752

26.7 Some Useful Real-Time Hardware Options754

26.8 Systems Design Considerations755

26.9 Design of Inspection Systems―the Status Quo757

26.10 System Optimization760

26.11 Concluding Remarks761

26.12 Bibliographical and Historical Notes763

26.12.1 General Background763

26.12.2 Developments Since 2000764

26.12.3 More Recent Developments765

CHAPTER 27 Epilogue―Perspectivesin Vision767

27.1 Introduction767

27.2 Parameters of Importance in Machine Vision768

27.3 Tradeoffs770

27.3.1 Some Important Tradeoffs770

27.3.2 Tradeoffs for Two-Stage Template Matching771

27.4 Moore's Law in Action772

27.5 Hardware,Algorithms,and Processes773

27.6 The Importance of Choice of Representation774

27.7 Past,Present,and Future775

27.8 Bibliographical and Historical Notes777

Appendix A Robust Statistics778

References796

Author Index845

Subject Index861