Research And Implementation Of PCB Welding Defect Detection Method Based On Image Processing

Based on image processing technology,research on surface soldering defect detection methods for printed circuit board(PCB)components processed by Surface Mounted Technology(SMT).It mainly focuses on four types of defects: lack of solder,bridging,crooked paste,and low tin.The specific work content is as follows:Research on image preprocessing methods.Combining the characteristics of the PCB image,the non-local mean filtering algorithm is used to filter out noise interference;the histogram is used to balance and stretch the image contrast;the Laplacian sharpening is used to improve the image clarity;the Canny operator is used for edge extraction.The circle detection technology is used to accurately locate the center of the positioning hole in the PCB image,combined with the data in the PCB design project to achieve the precise positioning of the components in the image,and then to achieve the extraction of the components and the solder joint area in the image;in the method of detecting the center of the positioning hole,The optimized use method of the Hough gradient method is proposed to ensure the accuracy of circle detection while improving the detection efficiency;the basic feature extraction method of the component and its solder joint area is studied;the matching degree of the component area of the test sample and the standard sample is used to determine whether the defect is Exist,then according to different defect surface characteristics,combined with decision tree rules to complete defect classification.Build a software and hardware platform,test the above methods through 200 PCB sample drawings,and analyze the experimental results to further optimize the defect recognition algorithm.In the end,the missed detection rate of the four types of defects is within 5%,which has certain practical application value.