机器视觉(或称计算机视觉)领域的研究博大精深，而且日新月异，对子具体视觉应用系统的设计人员和用户来说，该从何着手呢?本书是机器视觉领域的一本入门教材，详细介绍了基本概念，并辅以必要的数学知识，用较大篇幅来讲解如何在实际应用中实现和使用视觉算法，同时强调了技术的工程层面。本书有意省略了机器视觉中某些没有充分实际应用的理论。
　　本书可以作为高校相关专业的教材，也适合希望应用机器视觉来解决实际问题的各类人员阅读。

Ramesh Jain is currently a Professor of Electrical and Computer Engineer-ing, and Computer Science and Engineering at the University of California at San Diego. Before joining UCSD, he was a Professor of Electrical Engi-neering and Computer Science and the founding Director of the Artificial Intelligence Laboratory at the University of Michigan, Ann Arbor. He was also the founder and chairman of Imageware Inc. His current research in-terests are in multimedia information systems, image databases, machine vision, and intelligent systems.
　　Ramesh is a Fellow of IEEE, AAAI, and Society of Photo-Optical In-strumentation Engineers, and member of ACM, Pattern Recognition So-ciety, Cognitive Science Society, Optical Society of America, and Society of Manufacturing Engineers. He is currently the Editor-in-Chief of IEEE
Multimedia, and is on the editorial boards of Machine Vision and Applica-tions, Pattern Recognition, and Image and Vision Computing. He received his Ph.D. from IIT, Kharagpur, in 1975 and his B.E. from Nagpur Univer-sity in 1969.
　　Rangachar Kasturi joined Penn State University in 1982 after complet-ing his graduate studies at Texas Tech University (Ph.D. 1982 and M.S.E.E.1980). He received a B.E. (Electrical) degree from Bangalore University in 1968. His primary research focus in recent years has been in the area of Doc-ument Image Analysis (DIA). His group's main contribution has been the design of efficient algorithms to generate intelligent interpretations of engi-neering drawings and maps to facilitate automatic conversion from paper medium to computer databases. He is the Editor-in-Chief of IEEE Transac-tions on Pattern Analysis and Machine Intelligence. He was the managing editor of Machine Vision and Applications during 1993-94. He is a coau-thor of the tutorial texts, Computer Vision: Principles and Applications and Document Image Analysis, both published by IEEE CS Press, and a coeditor of the book Image Analysis Applications (Marcel Dekker, 1990).During 1987-90 he delivered lectures at many chapters of the IEEE Com-puter Society through its Distinguished Visitor Program. He has served the International Association for Pattern Recognition in various capacities.
　　Brian G. Schunck has worked for several years on the development of systems for machine vision and image processing. He was educated in computer science at the University of California, Irvine, where he received the B.S. magna cum laude in 1976. He studied electrical engineering, systems theory, and artificial intelligence at M.I.T., where he received the Master's
and E.E. degrees in 1979 for work on control algorithms for robotic manip ulators and the doctorate in 1983 for research on image flow. He was an assistant professor in the Department of Electrical Engineering and Computer Science and a member of the Artificial Intelligence Laboratory at the University of Michigan, Ann Arbor. Currently he is the Director of Vision
Software at Adept Technology.
　　Brian's current interests include statistical methods for machine vision and industrial inspection; contour, surface, and volume models forcomputer vision and medical image processing; structure and motion estimation for mobile robots; reverse engineering part models from range data; computer graphics; user interfaces; and marine navigation.
　　Brian Schunck is a member of the IEEE, ACM, the Society for Indus-trial and Applied Mathematics, the American Statistical Association, the American Society for Photogrammetry and Remote Sensing, the Society for Manufacturing Engineers, and the Society for Automotive Engineers.

This book grew out of our efforts to provide a balanced coverage of essential elements of machine vision systems to students in our undergraduate and early graduate classes. The field of machine vision, or computer vision, has been growing at a fast pace. The growth in this field, unlike most established fields, has been both in breadth and depth of concepts and techniques. To make the situation more confusing, the number of new applications has also been growing. Machine vision techniques are being applied in areas ranging from medical imaging to remote sensing, industrial inspection to document processing, and nanotechnology to multimedia databases.
　　As in most developing fields, not all aspects of machine vision that are of interest to active researchers are useful to the designers of a vision system for a specific application. A designer needs to know basic concepts and techniques to be successful in designing or evaluating a vision system for a particular application. It may not be necessary to know the latest, often controversial, results from leading research centers. On the other hand, the techniques learned by a designer should not be ephemeral.
　　This text is intended to provide a practical introduction to machine vi-sion. We made efforts to provide all of the details to allow vision algorithms to be used in practical applications. Intentionally omitted are theories of machine vision that do not appear to have sufficient practical applications at this time. We want this to be a useful introduction to machine vision rather than a state-of-the-art collection of research on machine vision.
　　The text is intended to be used in an introductory course in machine vision at the undergraduate or early graduate level and should be suitable for students with no prior knowledge of computer graphics or signal processing.
　　Students should have a working knowledge of mathematics through calculus of two variables, including matrices and linear spaces, and familiarity with basic probability theory, computer programming, and elementary data structures.　　Numerical and statistical methods and advanced algorithms are described as needed as well as material on geometry in two and three di-mensions. For some sections in the book, more mathematical background is needed. Such sections can be omitted by readers not interested in the rigorous formalization. We have made efforts to provide intuitive concepts,even for mathematical sections, that will help a reader understand the basic elements without the details.
　　An introductory text is based on material from several sources. This book also contains material from research papers, books, and other places.We have made no attempt to exhaustively list all original sources. We do provide some pointers to readers who are interested in exploring topics more deeply in each chapter. The references at the end of the book provide a list of sources that were directly used in the preparation of the book.
　　We strongly encourage readers to send any comments and corrections by mail to one of the authors or electronically to jain@ece.ucsd.edu.
　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　Ramesh Jain
Rangachar Kasturi
Brian G. Schunck

(美)Ramesh Jain,Rangachar Kasturi,Brian G.Schunck：Ramesh Jain: Ramesh Jain创建了密歇根大学的人工智能实验室，目前是加利福尼亚大学圣迭戈分校(UCSD)电气和计算机工程。计算机科学和工程系的教授。他的研究方向是多媒体信息系统。图像数据库。机器视觉和智能系统。他是《IEEE Multimedia》杂志的主编，《Machine Vision and Application》。《Pattern Recognition》和《Image and Vision Computing》杂志编委会成员，还是IEEE和AAAI的特别会员，ACM的会员。 　　
Rangachar Kasturi: Rangachar Kasturi 于得克萨斯技术大学获得博士学位之后到宾夕法尼亚州立大学执教。他的主要研究方向是文档图像分析(DIA)。他是《IEEE Transactions On Pattern Analysis and MachineIntelligence》杂志的主编。 　　
Brian G.Schunck: Brian G．Schunck于加州大学欧文分校获得学士学位，子麻省理工学院获硕士和博士学位。他目前是密歇根大学安阿伯分校电子工程与计算机科学系副教授，近年来一直从事机器视觉和图像处理系统的开发工作。

Preface
Acknowledgments
Introduction
1.1　Machine Vision
1.2　Relationships to Other Fields
1.3　Role of Knowledge
1.4　Image Geometry
1.4.1　Perspective Projection
1.4.2　Coordinate Systems
1.5　Sampling and Quantization
1.6　Image Definitions
1.7　Levels of Computation
1.7.1　Point Level
1.7.2　Local Level
1.7.3　Global Level
1.7.4　Object Level
1.8　Road Map
2　Binary Image Processing
2.1　Thresholding
2.2　Geometric Properties
2.2.1　Size
2.2.2　Position
2.2.3　Orientation
2.3　Projections
2.4　Run-Length Encoding
2.5　Binary Algorithms
2.5.1　 Definitions
2.5.2　 Component Labeling
2.5.3　 Size Filter
2.5.4　 Euler Number
2.5.5　 Region Boundary
2.5.6　 Area and Perimeter
2.5.7　 Compactness
2.5.8　 Distance Measures
2.5.9　 Distance Transforms
2.5.10　Medial Axis
2.5.11　Thinning
2.5.12　Expanding and Shrinking
2.6　Morphological Operators
2.7　Optical Character Recognition
3　Regions
3.1　Regions and Edges
3.2　Region Segmentation
3.2.1　Automatic Thresholding
3.2.2　Limitations of Histogram Methods
3.3　Region Representation
3.3.1　Array Representation
3.3.2　Hierarchical Representations
3.3.3　Region Characteristic-Based Representations
3.3.4　Data Structures for Segmentation
3.4　Split and Merge
3.4.1　Region Merging
3.4.2　Removing Weak Edges
3.4.3　Region Splitting
3.4.4　Split and Merge
3.5　Region Growing
4　Image Filtering
4.1　Histogram Modification
4.2　Linear Systems
4.3　Linear Filters
4.4　Median Filter
4.5　Gaussian Smoothing
4.5.1　Rotational Symmetry
4.5.2　Fourier Transform Property
4.5.3　Gaussian Separability
4.5.4　Cascading Gaussians
4.5.5　Designing Gaussian Filters
4.5.6　Discrete Ganssian Filters
5　Edge Detection
5.1　 Gradient
5.2　 Steps in Edge Detection
5.2.1　Roberts Operator
5.2.2　Sobel Operator
5.2.3　Prewitt Operator
5.2.4　Comparison
5.3　 Second Derivative Operators
5.3.1　Laplacian Operator
5.3.2　Second Directional Derivative
5.4　 Laplacian of Gaussian
5.5　 Image Approximation
5.6　 Gaussian Edge Detection
5.6.1　Canny Edge Detector
5.7　 Subpixel Location Estimation
5.8　 Edge Detector Performance
5.8.1　Methods for Evaluating Performance
5.8.2　Figure of Merit
5.9　 Sequential Methods
5.10　Line Detection
6　Contours
6.1　Geometry of Curves
6.2　Digital Curves
6.2.1　Chain Codes
6.2.2　Slope Representation
6.2.3　Slope Density Function
6.3　Curve Fitting
6.4　Polyline Representation
6.4.1　Polyline Splitting
6.4.2　Segment Merging
6.4.3　Split and Merge
6.4.4　Hop-Along Algorithm
6.5　Circular Arcs
6.6　Conic Sections
6.7　Spline Curves
6.8　Curve Approximation
6.8.1　Total Regression
6.8.2　Estimating Corners
6.8.3　Robust Regression
6.8.4　Hough Transform
6.9　Fourier Descriptors
7　Texture
7.1　Introduction
7.2　Statistical Methods of Texture Analysis
7.3　Structural Analysis of Ordered Texture
7.4　Model-Based Methods for Texture Analysis
7.5　Shape from Texture
8　Optics
8.1　Lens Equation
8.2　Image Resolution
8.3　Depth of Field
8.4　View Volume
8.5　Exposure
9　Shading
9.1　Image Irradiance
9.1.1　Illumination
9.1.2　Reflectance
9.2　Surface Orientation
9.3　The Reflectance Map
9.3.1　Diffuse Reflectance
9.4　Shape from Shading
9.5　Photometric Stereo
l0　Color
10.1　Color Physics
10.2　Color Terminology
10.3　Color Perception
10.4　Color Processing
10.5　Color Constancy
10.6　Discussion
11　Depth
11.1　Stereo Imaging
11.1.1.　Cameras in Arbitrary Position and Orientation
11.2　Stereo Matching
11.2.1　Edge Matching
11.2.2　Region Correlation
11.3　Shape from X
11.4　Range Imaging
11.4.1　Structured Lighting
11.4.2　Imaging Radar
11.5　Active Vision
12　Calibration
12.1　 Coordinate Systems
12.2　 Rigid Body Transformations
12.2.1　Rotation Matrices
12.2.2　Axis of Rotation
12.2.3　Unit Quaternions
12.3　 Absolute Orientation
12.4　 Relative Orientation
12.5　 Rectification
12.6　 Depth from Binocular Stereo
12.7　 Absolute Orientation with Scale
12.8　 Exterior Orientation
12.8.1　Calibration Example
12.9　 Interior Orientation
12.10　Camera Calibration
12.10.1　Simple Method for Camera Calibration
12.10.2　Affine Method for Camera Calibration
12.10.3　Nonlinear Method for Camera Calibration
12.11　Binocular Stereo Calibration
12.12　Active Triangulation
12.13　Robust Methods
12.14　Conclusions
13　Curves and Surfaces
13.1　Fields
13.2　Geometry of Curves
13.3　Geometry of Surfaces
13.3.1　Planes
13.3.2　Differential Geometry
13.4　Curve Representations
13.4.1　Cubic Spline Curves
13.5　Surface Representations
13.5.1　Polygonal Meshes
13.5.2　Surface Patches
13.5.3　Tensor-Product Surfaces
13.6　Surface Interpolation
13.6.1　Triangular Mesh Interpolation
13.6.2　Bilinear Interpolation
13.6.3　Robust Interpolation
13.7　Surface Approximation
13.7.1　Regression Splines
13.7.2　Variational Methods
13.7.3　Weighted Spline Approximation
13.8　Surface Segmentation
13.8.1　Initial Segmentation
13.8.2　Extending Surface Patches
13.9　Surface Registration
14　Dynamic Vision
14.1　Change Detection
14.1.1　Difference Pictures
14.1.2　Static Segmentation and Matching
14.2　Segmentation Using Motion
14.2.1　Time-Varying Edge Detection
14.2.2　Stationary Camera
14.3　Motion Correspondence
14.4　Image Flow
14.4.1　Computing Image Flow
14.4.2　Feature-Based Methods
14.4.3　Gradient-Based Methods
14.4.4　Variational Methods for Image Flow
14.4.5　Robust Computation of Image Flow
14.4.6　Information in Image Flow
14.5　Segmentation Using a Moving Camera
14.5.1　Ego-Motion Complex Log Mapping
14.5.2　Depth Determination
14.6　Tracking
14.6.1　Deviation Function for Path Coherence
14.6.2　Path Coherence Function
14.6.3　Path Coherence in the Presence of Occlusion
14.6.4　Modified Greedy Exchange Algorithm
14.7　Shape from Motion Object Recognition
15.1　System Components
15.2　Complexity of Object Recognition
15.3　Object Representation
15.3.1　Observer-Centered Representations
15.3.2　Object-Centered Representations
15.4　Feature Detection
15.5　Recognition Strategies
15.5.1　Classification
15.5.2　Matching
15.5.3　Feature Indexing
15.6　Verification
15.6.1　Template Matching
15.6.2　Morphological Approach
15.6.3　Symbolic
15.6.4　Analogical Methods
A　Mathematical Concepts
A.1　Analytic Geometry
A.2　Linear Algebra
A.3　Variational Calculus
A.4　Numerical Methods
B　Statistical Methods
B.1　Measurement Errors
B.2　Error Distributions
B.3　Linear Regression
B.4　Nonlinear Regression
C　Programming Techniques
C.1　Image Descriptors
C.2　Mapping Operators
C.3　Image File Formats
Bibliography
Index