Research On Some Key Technologies Of PCB Defect Detection System Based On Machine Vision

With the development of society and economy,the application of printed circuit board(PCB)is becoming more and more extensive,and the quality requirements are increasing.Aiming at the disadvantages of manual visual inspection and the costliness of foreign inspection equipment,this paper designed and builded a machine visionbased PCB defect detection system.Combined machine vision technology with digital image processing technology key technologies such as image orientation rectification were studied,device missing part detection,and chip polarity defect detection.The main research contents completed are as follows:(1)Analyze the research and application status of PCB defect detection related technologies based on machine vision at home and abroad.Combined with the structural characteristics of the inspection system and the shortcomings of defect detection technology,a PCB defect detection platform using machine vision technology and image processing technology is build;(2)Aiming at the problems of low accuracy of traditional image orientation rectification methods,a PCB image orientation rectification method based on multidirectional improved Sobel operator is proposed.This method uses multi-directional improved Sobel operator to extract the edge information of the G channel of the PCB image;binarizes the edge image,combines the median filtering to extract the image outline,and realizes the PCB region localization;finally,the whole image is corrected by affine transformation in the region of interest.Through comparative experimental analysis,the PCB image positioning correction algorithm proposed in this paper is superior to the existing main methods,which effectively improves the accuracy of image positioning correction;(3)In view of the lack of accuracy of the current template matching method and defect detection method applied to PCB defect detection systems,this paper proposes a multi-level PCB defect detection method based on improved template matching and image difference methods.First,the improved template matching proposed in this paper is used to capture the image to be tested.For missing parts in PCB boards,a multi-level binary detection method based on image difference method is proposed to implement PCB image defect detection.Combined with the experimental platform test,it is found that the method can accurately extract the defect area.Compared with existing methods,this method has better stability and higher accuracy;(4)Aiming at the chip polarity detection problem in PCB board defect detection,a chip polarity detection method combining regions of interest,improved template matching,and pixel addition is proposed to achieve chip polarity detection.First,determine the region of interest through prior knowledge,and use template matching to quickly and accurately position the chip.Then introduce histogram matching and median filtering to preprocess the chip image.Finally,use image binarization and pixel addition chip polarity judgment.Experimental results show that the proposed method has higher accuracy in chip positioning and polarity detection.Through experimental comparison and analysis,it is found that the algorithm proposed in this paper has higher effectiveness and reliability,which lays the foundation for subsequent function expansion and defect detection in the site environment.