Contents Chapter 1Introduction 1.1Machine vision technology 1.1.1The development of machine vision technology 1.1.2The application of machine vision technology 1.1.3Composition of machine vision system 1.1.4Advantages of machine vision system 1.2Research and applications of metal surface inspection 1.3Surface defect detection and identification algorithms 1.4Challenges and development 1.5The main content and basic structure References Chapter 2Composition of Online Detection System 2.1Imaging device 2.1.1Industrial cameras 2.1.2Camera imaging model 2.1.3Lens 2.1.4Main parameters and calculations of imaging devices 2.2Light source 2.2.1Incandescent lamp 2.2.2Halogen lamp 2.2.3Gas discharge lamp 2.2.4Light��emitting diode 2.2.5Laser light source 2.3Data acquisition controller 2.4Mechanical structure and supporting facilities 2.5Data processing and computing system 2.5.1Graphical user interface 2.5.2Algorithms 2.5.3System architecture References Chapter 3Image Processing and Recognition Algorithms 3.1A review of digital image processing 3.1.1Image and digital image 3.1.2Image processing technology 3.1.3Image engineering 3.1.4Surface defect detection algorithms 3.2Image restoration 3.2.1Theoretical model of image restoration 3.2.2Spatial filtering 3.3Feature extraction 3.3.1Overview 3.3.2Geometric feature extraction 3.3.3Gray level histogram feature extraction 3.3.4Image texture feature extraction 3.3.5Feature point extraction and description 3.3.6Deep learning feature extraction 3.4Image classification 3.4.1Deep convolutional neural networks 3.4.2Classic deep convolutional neural networks 3.4.3Deep convolutional neural networks with attention mechanism 3.4.4Support Vector Machine References Chapter 4Multiple Information Fusion for Defect Detection 4.1Multi��information fusion 4.1.1Concept of information fusion 4.1.2Hierarchical structure of information fusion 4.1.3Overview of information fusion algorithms 4.2Deep 3D object detection for point cloud data 4.2.13D detection techniques 4.2.2RGB��D 3D detection techniques 4.33D detection of surface defects in high��temperature castings 4.3.1Types and characteristics of surface defects in high��temperature castings 4.3.23D shape reconstruction 4.3.3Overview of high��temperature casting 3D inspection system 4.3.4Design of high temperature casting billet 3D detection system 4.3.5Hardware selection for high��temperature casting 3D inspection system 4.3.6Imaging scheme for 3D inspection system 4.3.7Algorithm for fusion of gray��level and depth information in high��temperature castings References Chapter 5Deployment of Online Detection Algorithms 5.1Real��time requirements of surface online inspection techniques 5.1.1Online inspection techniques 5.1.2Conventional surface defect detection methods 5.1.3Real��time online surface inspection techniques 5.2Algorithm multi��threading acceleration 5.2.1Introduction to threads 5.2.2Introduction to multi��threading 5.2.3Introduction to multi��threading in Python 5.2.4Thread synchronization in python 5.2.5Global interpreter lock 5.3Algorithm multi��processing acceleration 5.3.1Multi��programming techniques 5.3.2Process scheduling 5.3.3Process state 5.3.4Python multi��processing 5.3.5Multi��process realization 5.4GPU acceleration of algorithms 5.4.1Training and deployment of deep learning 5.4.2Optimization principles of TensorRT 5.4.3Optimization steps of TensorRT 5.4.4GPU parallel acceleration 5.4.5NVIDIA GPU acceleration application case study 5.4.6Huawei Atlas GPU acceleration application case study References Chapter 6High��speed Wire Surface Online Inspection System 6.1The demand for online surface inspection of high��speed wire 6.1.1Background of online surface inspection for high��speed wire 6.1.2Requirements for online surface inspection of high��speed wire 6.2High��speed wire surface imaging system and image characteristics 6.2.1High��speed wire surface imaging system 6.2.2Characteristics of the images 6.3Correction of high��speed wire surface images 6.3.1Reasons for correction 6.3.2Basic principles and bottlenecks of correction 6.3.3Advantages and disadvantages of different correction methods 6.4Principles of defect detection algorithms for high��speed wire surface images 6.4.1Experimental data 6.4.2Data augmentation 6.4.3K��means 6.4.4DIoU��NMS 6.4.5Evaluation metrics for object detection 6.4.6Model training 6.5Deployment of defect detection algorithms for high��speed wire surface 6.5.1Introduction of hardware 6.5.2Technical introduction 6.5.3Software deployment 6.5.4Deployment effectiveness References