Research On PCB Surface Defect Detection Technology Based On Machine Vision

In recent years,with the continuous development of PCB printing technology and the continuous improvement of process level,PCB products are also developing rapidly towards ultra-thin,small components,high density and fine spacing.The assembly density of components on PCB is increasing,the line width,spacing and pad of PCB are becoming smaller and smaller,and the number of composite layers is increasing,which makes the quality detection of PCB such as short circuit,circuit break,breakage and copper slag a more difficult task.Traditional manual detection methods are easy to miss detection,slow detection speed and long detection time,which can not meet the needs of production.Machine vision-based surface defect detection technology has been widely used in various fields of visual inspection.However,PCB board surface defect detection technology still faces problems such as slow detection speed,low recognition accuracy and inefficient use of PCB board damaged in a reasonable range.In view of the above problems,this paper designs a PCB defect visual detection system,the specific research contents are as follows:(1)PCB image segmentation algorithm is studied,and the bimodal threshold method,OTSU threshold method and SVM method are compared and analyzed,it is concluded that SVM method is more suitable for PCB board studied in this paper.The experimental results show that compared with the bimodal threshold method and OTSU threshold method,the SVM method can preserve the integrity of the pixels well,and can recognize and segment the small area more accurately.(2)PCB image registration algorithm is studied,the image registration based on feature and mutual information is compared and analyzed,it is concluded that image registration based on mutual information is more suitable for PCB board studied in this paper.The experimental results show that the registration based on mutual information is more accurate and faster than that based on feature.(3)Aiming at the defects of PCB circuit board,a defect detection method based on gray histogram is proposed.Firstly,the image of the defect is corrected,and the defect area is obtained by image difference between the template image and the corrected image.Then Roberts edge detection is applied to defect area to get defect area contour.Finally,the defect type is identified by calculating the gray histogram of the defect area contour.The experimental results show that the algorithm can accurately detect and identify defects such as short circuit,circuit break,breakage and copper slag,with high recognition accuracy and fast detection speed.(4)A distance detection method is proposed for calculating the damage rate of the damaged defect area.The method calculates the breakage rate by extracting the conductor skeleton and calculating the maximum distance from the point on the skeleton to the edge of the conductor.Experiments show that the method can accurately calculate the breakage rate.(5)PCB defect detection system is designed,including hardware selection,such as camera,lens,light source,etc.The testing software system is compiled.After field test,the whole system is stable and effective.