第1页 / 共802页
第2页 / 共802页
第3页 / 共802页
第4页 / 共802页
第5页 / 共802页
第6页 / 共802页
第7页 / 共802页
第8页 / 共802页
机器视觉自动视觉检测理论、实践与应用（英文原版）.pdf
Preface
Contents
Chapter 1 Introduction
1 Introduction
1.1 Visual inspection
1.2 Optical capturing of test objects
1.3 Formation and definition of an image
1.4 Machine vision
1.5 Practical approach for performing machine vision projects
1.6 Bibliography
Part I Image Acquisition
Chapter 2 Light
2 Light
2.1 The phenomenon of light
2.1.1 The electromagnetic spectrum
2.2 Light as an electromagnetic wave
2.2.1 Maxwell’s equations
2.2.2 Polarization
2.2.3 Huygens’ principle
2.2.4 Coherence
2.2.5 Interference
2.2.6 Diffraction
2.2.7 Speckle
2.3 Light as a quantum phenomenon
2.4 The ray model of geometrical optics
2.5 Summary
2.6 Interaction of light and matter
2.6.1 Absorption
2.6.2 The law of reflection
2.6.3 The law of refraction
2.6.4 Scattering
2.6.5 The Fresnel coefficients for reflection and transmission
2.6.6 Electromagnetic waves in conductive media
2.7 Light sources
2.7.1 Thermal radiators
2.7.2 Gas-discharge lamps
2.7.3 Light-emitting diodes
2.7.4 Laser
2.7.5 Summary
2.8 Bibliography
Chapter 3 Optical Imaging
3 Optical Imaging
3.1 Introduction
3.2 Imaging with a pinhole camera, central projection
3.3 The camera model and camera calibration
3.4 Optical imaging using a single lens
3.4.1 The paraxial approximation and Gaussian optics
3.4.2 Thin lens equation
3.4.3 Bundle limitation
3.4.4 Depth of field
3.4.5 Telecentric imaging
3.4.6 Perspective
3.4.7 Imaging of tilted planes
3.4.8 Aberrations
3.5 Optical instruments with several lenses
3.5.1 The projector
3.5.2 The microscope
3.6 Bibliography
Chapter 4 Radiometry
4 Radiometry
4.1 Radiometric quantities
4.2 The light field of a test object
4.3 The bidirectional reflectance distribution function (BRDF)
4.3.1 BRDF and scattered light
4.4 Formation of image values
4.4.1 Application to a thin lens
4.5 Bibliography
Chapter 5 Color
5 Color
5.1 Photometry
5.2 Color perception and color spaces
5.2.1 Color perception of the human eye
5.2.2 Color mixing
5.2.3 CIE color spaces
5.2.4 Spectrophotometry for color measurement and color distance computation
5.2.5 Color order systems
5.2.6 Other color spaces
5.3 Filters
5.4 Acquisition and processing of color images
5.5 Bibliography
Chapter 6 Sensors for Image Acquisition
6 Sensors for Image Acquisition
6.1 Point, line and area sensors
6.2 Image tube cameras
6.3 Photomultipliers
6.3.1 Image intensifiers
6.4 Photodiodes
6.5 Position sensitive detectors (PSD)
6.6 Charge-coupled device (CCD)
6.7 Complementary metal-oxide-semiconductor (CMOS) sensors
6.8 Line-scan cameras
6.9 Color sensors and color cameras
6.10 Infrared cameras
6.10.1 Bolometer cameras
6.10.2 Infrared quantum detector cameras
6.11 Quality criteria for image sensors
6.12 Bibliography
Chapter 7 Methods of Image Acquisition
7 Methods of Image Acquisition
7.1 Introduction
7.2 Measuring optical properties
7.2.1 Measurement of the complex index of refraction
7.2.2 Fluorescence
7.2.3 Methods for measuring the reflectance
7.2.4 Spectral sensors
7.2.5 Light scattering methods and the inspection of surface roughness
7.3 3D shape capturing
7.3.1 Triangulation (point-by-point scanning)
7.3.2 Light-section methods (line scanning)
7.3.3 The measurement uncertainty of triangulation
7.3.4 Structured illumination
7.3.5 Deflectometry
7.3.6 The moiré method
7.3.7 Final remark on structured illumination
7.3.8 Stereo images
7.3.9 Light-field cameras
7.3.10 Silhouette capturing
7.3.11 Shape from shading
7.3.12 Autofocus sensors
7.3.13 Confocal microscopy
7.3.14 Confocal chromatic triangulation
7.3.15 Time-of-flight sensors
7.3.16 Phase-based methods
7.4 Capturing interior object structures
7.4.1 Thermography
7.4.2 Imaging using X-rays
7.4.3 Optical coherence tomography
7.4.4 Schlieren imaging and schlieren tomography
7.4.5 Image acquisition using terahertz radiation
7.4.6 Photoelasticity
7.5 Special image acquisition methods
7.5.1 Image acquisition systems with variable illumination direction
7.5.2 Endoscopy
7.6 Universal principles
7.6.1 Suppression of extraneous light
7.6.2 Inverse illumination
7.7 Summary
7.8 Bibliography
Part II Image Processing
Chapter 8 Image Signals
8 Image Signals
8.1 Mathematical model of image signals
8.2 Systems and signals
8.2.1 System characteristics
8.2.2 The Dirac delta function
8.2.3 Convolution
8.3 The Fourier transform
8.3.1 The one-dimensional Fourier transform
8.3.2 The one-dimensional sampling theorem
8.3.3 The discrete Fourier transform (DFT)
8.3.4 The two-dimensional Fourier transform
8.3.5 Dirac delta functions in two-dimensional space
8.3.6 The two-dimensional Heaviside function
8.3.7 Sampling of two-dimensional signals
8.3.8 Sampling theorem for two-dimensional signals
8.3.9 The two-dimensional DFT
8.4 Examples of use concerning system theory and the Fourier transform
8.5 Image signals as stochastic processes
8.5.1 Moments of stochastic processes
8.5.2 Stationarity and ergodicity
8.5.3 Passing a stochastic process through an LSI system
8.6 Quantization
8.6.1 Optimal quantization
8.6.2 The quantization theorem
8.6.3 Modeling of the quantization
8.7 The Karhunen–Loève transform
8.7.1 Definition of the Karhunen–Loève transform
8.7.2 Properties of the Karhunen–Loève transform
8.7.3 Examples of application of the Karhunen–Loève transform
8.8 Bibliography
Chapter 9 Preprocessing and Image Enhancement
9 Preprocessing and Image Enhancement
9.1 Simple image enhancement methods
9.1.1 Contrast adjustment by histogram stretching
9.1.2 Histogram manipulation
9.1.3 Pseudo-color and false-color images
9.1.4 Image sharpening
9.2 Reduction of systematic errors
9.2.1 Geometric rectification
9.2.2 Suppression of inhomogeneities
9.3 Attenuation of random disturbances
9.3.1 Linear filters
9.3.2 Noise reduction using nonlinear filters
9.4 Image registration
9.5 Bibliography
Chapter 10 Image Restoration
10 Image Restoration
10.1 Signal model
10.2 Inverse filter
10.3 The Wiener filter
10.4 The geometric mean filter
10.5 Optimal constraint filter
10.6 Restoration problems in matrix notation
10.7 Restoration for participating media
10.8 Spatially-varying image restoration
10.9 Bibliography
Chapter 11 Segmentation
11 Segmentation
11.1 Region-based segmentation
11.1.1 Segmentation by feature-based classification
11.1.2 Region growing methods
11.2 Edge-oriented methods
11.2.1 Gradient filters
11.2.2 Edge detection using the second derivative
11.2.3 The watershed transformation
11.3 Diffusion filters
11.3.1 Linear, homogeneous, isotropic image diffusion
11.3.2 Linear, inhomogeneous, isotropic image diffusion
11.3.3 Nonlinear, inhomogeneous, isotropic image diffusion
11.3.4 Nonlinear, inhomogeneous, anisotropic image diffusion
11.4 Active contours
11.4.1 Gradient vector flow
11.4.2 Vector field convolution
11.5 Segmentation according to Mumford and Shah
11.6 Segmentation using graph cut methods
11.7 Bibliography
Chapter 12 Morphological Image Processing
12 Morphological Image Processing
12.1 Binary morphology
12.1.1 Point sets and structuring elements
12.1.2 Erosion and dilation
12.1.3 Opening and closing
12.1.4 Border extraction
12.1.5 Region filling
12.1.6 Component labeling and connected component analysis
12.1.7 The hit-or-miss operator
12.1.8 Skeletonization
12.1.9 Pruning
12.2 Gray-scale morphology
12.2.1 The point set of a gray-scale image
12.2.2 Erosion and dilation
12.2.3 Opening and closing
12.2.4 Edge detection
12.3 Bibliography
Chapter 13 Texture Analysis
13 Texture Analysis
13.1 Types of textures
13.1.1 Structural texture type
13.1.2 Structural-statistical texture type
13.1.3 Statistical texture type
13.2 Visual inspection tasks regarding textures
13.3 Model-based texture analysis
13.3.1 Analysis of structural textures
13.3.2 Analysis of structural-statistical textures
13.3.3 Autoregressive models for analyzing statistical textures
13.3.4 Separation of line textures
13.4 Feature-based texture analysis
13.4.1 Basic statistical texture features
13.4.2 Co-occurrence matrix
13.4.3 Histogram of oriented gradients
13.4.4 Run-length analysis
13.4.5 Laws’ texture energy measures
13.4.6 Local binary patterns
13.5 Bibliography
Chapter 14 Detection
14 Detection
14.1 Detection of known objects by linear filters
14.1.1 Unknown background
14.1.2 White noise as background
14.1.3 Correlated, weakly stationary noise as background
14.1.4 Discrete formulation of the matched filter
14.2 Detection of unknown objects (defects)
14.3 Detection of straight lines
14.3.1 The Radon transform
14.3.2 Detection of line-shaped structures
14.3.3 The Hough transform for the detection of lines
14.3.4 The Hough transform for the detection of curves
14.3.5 The generalized Hough transform
14.3.6 Implicit shape models
14.4 Corner detection
14.5 Bibliography
Chapter 15 Image Pyramids, theWavelet Transform and
Multiresolution Analysis
15 Image Pyramids, the Wavelet Transform and Multiresolution Analysis
15.1 Image pyramids
15.1.1 Gaussian pyramid
15.1.2 Laplacian pyramid
15.1.3 Pyramid linking
15.2 Wavelets
15.2.1 Continuous wavelet transform
15.2.2 Discretization of the wavelet transform
15.3 Multiresolution analysis
15.4 The fast wavelet transform
15.5 The two-dimensional wavelet transform
15.6 Scale-invariant features
15.7 Bibliography
Part III Appendix
Appendix A Mathematical Foundations
A Mathematical Foundations
A.1 The intercept theorem
A.2 Inverse problems
A.3 Bibliography
Appendix B The Fourier Transform
B The Fourier Transform
B.1 The one-dimensional Fourier transform
B.1.1 Definition
B.1.2 Properties and characteristics
B.1.3 Correspondences
B.2 The n-dimensional Fourier transform
B.2.1 Definition
B.2.2 Correspondences of the two-dimensional Fourier transform
B.3 The discrete Fourier transform
List of Symbols
List of Abbreviations
Index
Machine Vision
Jürgen Beyerer • Fernando Puente León
Christian Frese
Machine Vision
Automated Visual Inspection:
Theory, Practice and Applications
123
Christian Frese
Fraunhofer-Institut für Optronik,
Systemtechnik und Bildauswertung
Karlsruhe, Germany
Jürgen Beyerer
Fraunhofer-Institut für Optronik,
Systemtechnik und Bildauswertung
and Karlsruhe Institute of Technology
Karlsruhe, Germany
Fernando Puente León
Karlsruhe Institute of Technology
Karlsruhe, Germany
Translator: Johannes Meyer
ISBN 978-3-662-47793-9
DOI 10.1007/978-3-662-47794-6
ISBN 978-3-662-47794-6 (eBook)
Library of Congress Control Number: 2015947141
Springer
© Springer-Verlag Berlin Heidelberg 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifi cally the rights of translation, reprinting, reuse of illustrations, recita-
tion, broadcasting, reproduction on microfi lms or in any other physical way, and transmission or in-
formation storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publica-
tion does not imply, even in the absence of a specifi c statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made.
Printed on acid-free paper
Springer is a brand of Springer Fachmedien Wiesbaden
Springer Berlin Heidelberg is part of Springer Science+Business Media
(www.springer.com)
Dedicated to Professor Dr.-Ing. Franz Mesch
VII
Preface
Machine vision and automated visual inspection are domains of automation technology
with a steadily increasing economical relevance. Although the related industry is notably ex-
panding since the past two decades, only a part of today’s visual inspection tasks have been
automated. This is why there is a great potential for economization in high income countries
which may lead to both reduced costs and increased quality of the produced goods.
As visual perception is the main human sensation, the automation of visual inspection is
somehow fascinating—at least for the authors of this book. When talking about the automa-
tion of visual inspection one might easily think that it cannot be that hard to teach a technical
visual inspection system to perceive what a human can easily see with only a glimpse. Ac-
tually, users often state: ‘As humans can see that instantly, it can’t be that hard to achieve the
same using a machine’. The answer is not easy at all but it depends on the individual case:
for humans, some things represent easy tasks which however are difficult to automate—on
the contrary, many other things can be done more precisely and particularly more reliably
by machines, if an automation is possible.
Automated visual inspection is a complex and multi-disciplinary topic involving optics,
mechanical and electrical engineering, mathematics and computer science. Systems for au-
tomated visual inspection are usually more or less complex mechatronic systems, which can
only achieve the requested performance in an economical way, if all the necessary disciplines
collaborate.
Everything starts with a visual inspection task that is to be carried out using an auto-
mated approach. In this context, image acquisition plays an important role: loss of informa-
tion during that step can hardly be compensated during later image processing steps. The
success of a visual inspection solution depends heavily on the quality of this first step.
Fortunately, when designing an automated visual inspection system, one usually has the
benefit of several degrees of freedom, in order to obtain image data with sufficient quality
and significance. This is why the suitability of the image acquisition for a given problem at
least partly depends on the engineer. In order to exploit those degrees of freedom at the best,
this book draws particular attention to image acquisition and the acquisition constellation,
consisting of the test object, the illumination and the acquisition system.
This book has the ambition to thoroughly introduce the reader into the terms of auto-
mated visual inspection. For this purpose, the Chapters 2 to 6 of the book’s first part deal
with the physics of image formation and the required optical principles and techniques in
an adequately extensive way. Based on that foundations, image acquisition for automated
visual inspection will be treated in Chapter 7. In this key chapter, a multitude of different
techniques for image acquisition will be explained in a systematic way, as well as important
hints and tricks will be shown which are indispensable for a good visual inspection system.
In order to enable automated analysis of images in a computer, the analog image signals
have to be transformed into digital signals. The underlying theory of signal processing and
the effects of local sampling and quantization will be extensively discussed, especially in
terms of system theory. Among others, Chapter 8 is devoted to the basics of digital process-
ing of analog signals and prepares the reader for the second part of the book, which focuses
on image analysis. Chapters 9 to 15 cover methods, which form the individual steps leading
to a final inspection result based on the acquired image data.
VIII
Preface
The depth of the explanations of all covered subjects is chosen to provide the reader with
insight into the respective motivation and backgrounds. No facts are supposed ‘to appear
from nowhere’; the underlying concepts should be thoroughly understood. Some theorems
however will not be proven in a strict mathematical way. In fact, there will be sketches of the
proofs, which will present their essential idea and help to understand important concepts.
For an application-oriented reader who is nevertheless interested in what happens behind
the scenes, consciously omitting technically flawlessly led proofs increases the book’s read-
ability and leads to a handy amount of pages.
The book on hand is partially based on lectures held by the author J. Beyerer at the Karls-
ruhe Institute of Technology (KIT, formerly University of Karlsruhe) since 1994 and by the
author F. Puente León, initially at Technische Universität München (TUM) since 2003 and at
KIT since 2008. It addresses itself to students studying in the fields of engineering science,
computer science, physics and mathematics. As all needed concepts and methods are intro-
duced in a sufficiently exhaustive way, it should be possible for advanced bachelor students
to clearly understand the presented content. Furthermore, scientists, PhD students and es-
pecially master students dealing with automated visual inspection can profit from reading
the book as its topics are appropriately elaborated.
Besides theory, practice is not missed out. The authors’ industrial experience, which is in-
corporated into many topics of the book, brings benefits even to practically oriented readers
who seek for robust and economic solutions for concrete visual inspection tasks. Nonethe-
less, the book does not loose itself into superficial recipes but yields enough substance for a
deep understanding of the presented content.
The authors want to particularly thank their following colleagues for supporting them in
the creation of this book:
Dr. Ulrich Breitmeier (Breitmeier Messtechnik GmbH, Ettlingen, Germany) for an exam-
ple image of a cylindric scanner
Dr. Michael Fried (University of Erlangen, Germany) for an example of the Mumford-
Shah-Method
Dr. Jan Horn (Department of Measurement and Control, Karlsruhe Institute of Technol-
ogy) for an image of camera based velocity measurement
Dr. Udo Netzelmann and Dr. Günter Walle (Fraunhofer IZFP, Saarbrücken, Germany) for
example images of the impulse thermography
Arne Nowak (Fraunhofer IIS, Erlangen, Germany) for example images acquired with the
POLKA camera
Dirk Nüßler (Fraunhofer FHR, Wachtberg, Germany) for example images of inspection
using terahertz radiation
Prof. Dr. Wolfgang Osten (University of Stuttgart, Germany) for images of interferometric
methods
Prof. Dr. Jerry L. Prince (Johns Hopkins University, Baltimore, USA) for examples of ac-
tive contours
Dr. Andreas Purde (Institute for Measurement Systems and Sensor Technology, TUM,
Germany) for an example for the pyramid linking method
Dr. Anna Remelli and Dr. Claudio Sedazzari (Opto Engineering, Mantova, Italy) for ex-
amples of the hypercentric perspective
Dr. Norman Uhlmann (Fraunhofer EZRT, Fürth, Germany) for example images of X-ray
inspection
Preface
IX
Bernhard Schmitt M.A. (ONUK, Karlsruhe, Germany) for an aerial photograph of the
Karlsruhe Palace
Dirk vom Stein and Thomas Winkel (Inspectomation GmbH, Mannheim, Germany) for
example images showing the inspection of casting cores and brake discs as well as tele-
centric images
Dr. Marco Kruse (Institute of Industrial Information Technology IIIT, Karlsruhe Institute
of Technology KIT) for the image of the checkerboard-shadow illusion as well as for ex-
ample images concerning the restoration of uniform motion blur
Mario Lietz (IIIT, KIT) for example images showing the division by a reference image
Dr. Ioana Ghe¸ta (Vision and Fusion Laboratory IES, Institute for Anthropomatics, Karls-
ruhe Institute of Technology) for examples concerning the analysis of spectral image se-
ries
Dr. Robin Gruna (IES, KIT) for images recorded with inverse illumination
Dr. Matthias Michelsburg (IIIT, KIT) and Dr. Robin Gruna for hyperspectral images of
food
Thomas Stephan (IES, KIT) for images recorded with a light-field camera and for exam-
ples showing the restoration of participating media
Dr. Matthias Hartrumpf (Fraunhofer Institute for Optronics, System Technologies and
Image Exploitation IOSB, Karlsruhe, Germany) for example images showing photoelas-
ticity and the inspection of glass preforms
Prof. Dr. Michael Heizmann (Fraunhofer IOSB) for examples of texture analysis as well
as shape from shading
Christian Negara (Fraunhofer IOSB) for examples of the graph cut method
Johannes Pallauf (IIIT, KIT) for the example of image restoration for uniform motion blur
Günter Saur and Wolfgang Roller (Fraunhofer IOSB) for supplying TerraSAR-X images
Günter Struck and Dr. Kai-Uwe Vieth (Fraunhofer IOSB) for examples of fluorescence
spectroscopy
Chen-Ko Sung (Fraunhofer IOSB) for example images of inspection based on flatbed
scanners
Dr. Stefan Werling (Fraunhofer IOSB) for deflectometric images
Dr. Alexander Schwarz and Martina Richter (Fraunhofer IOSB) for images showing an
example BRDF measurement
Dr. Miro Taphanel (IES, KIT) for images concerning the CCT sensor
Johannes Meyer (IES, KIT) for example images acquired with a Schlieren setup
Dr. Yaokun Zhang (IPR, KIT) and Johannes Meyer (IES, KIT) for example OCT images
Special thanks go to Johannes Meyer (IES, KIT) for the incorporation of numerous exten-
sions into the German version of this book (‘Automatische Sichtprüfung’) and for translat-
ing it into English. A big contribution to the contents presented in this book has been made
especially by Dr. Stefan Werling, Dr. Christoph Lindner, Dr. Ana Pérez Grassi, Dr. Robin
Gruna, Sebastian Höfer and Dr. Michael Teutsch by assisting the mentioned lectures. The
authors would like to also thank all the students who were involved in creating images and
diagrams. In addition, numerous students have sent in valuable suggestions for improving
the book’s didactic component.
Special thanks are directed to Andrey Belkin, Yvonne Fischer, Peter Frühberger, Dr.
Robin Gruna, Jan Hendrik Hammer, Pilar Hernández Mesa, Christian Herrmann, Sebas-
tian Höfer, Chettapong Janya-Anurak, Mahsa Mohammadi Kaji, Dr. Marco Kruse, Achim
相关推荐
2023年江西萍乡中考道德与法治真题及答案.doc
2012年重庆南川中考生物真题及答案.doc
2013年江西师范大学地理学综合及文艺理论基础考研真题.doc
2020年四川甘孜小升初语文真题及答案I卷.doc
2020年注册岩土工程师专业基础考试真题及答案.doc
2023-2024学年福建省厦门市九年级上学期数学月考试题及答案.doc
2021-2022学年辽宁省沈阳市大东区九年级上学期语文期末试题及答案.doc
2022-2023学年北京东城区初三第一学期物理期末试卷及答案.doc
2018上半年江西教师资格初中地理学科知识与教学能力真题及答案.doc
2012年河北国家公务员申论考试真题及答案-省级.doc
2020-2021学年江苏省扬州市江都区邵樊片九年级上学期数学第一次质量检测试题及答案.doc
2022下半年黑龙江教师资格证中学综合素质真题及答案.doc
