Research On The Defect Detection Method Of Flexible Circuit Board Gold Finger

Flexible circuit board(Flexible Printed Circuit,referred to as FPC)processing process has dozens of processes,will produce some defects,effective defect detection method is to ensure the quality of the FPC premise,for the FPC defect detection of non-reference method detection efficiency is low,and the reference method matching accuracy is poor,slow problem,this paper to the mobile phone FPC gold finger as the research object,carried out the FPC gold finger defect online detection method research,the specific research content is as follows:1)Firstly,the FPC image acquisition system was built,and then the study on the characteristics of defects such as short circuit,breakage,scratch,foreign matter,contamination and leakage plating was carried out by combining the structural characteristics of the single column of the gold finger.Combined with the geometry and image tone characteristics of FPC gold finger defects,the FPC defects were classified into two categories:I.Short-circuit and breakage defects;II.Scratch,contamination,foreign matter and leakage plating defects.In order to better segment FPC gold fingers and background areas,research on pepper noise removal methods based on mean filtering,Gaussian filtering and median filtering algorithms was carried out to compare and analyze the effect on pepper noise removal,pointing out that median filtering can remove pepper noise well;research on image segmentation methods based on threshold,boundary and region was carried out to compare and analyze the accuracy of segmentation results,pointing out that the threshold-based The maximum inter-class variance method can better segment the FPC images.2)For FPC gold finger short circuit and breakage defects,in order to improve the detection efficiency,the research of FPC defect detection method based on ZS skeleton extraction is carried out,and on the basis of ZS refinement algorithm,nine additional deletion templates are added to improve the ZS refinement algorithm to realize the single pixel skeleton extraction of gold finger image,to obtain the endpoint and node position information of gold finger skeleton,and to achieve the detection of FPC gold finger short circuit and breakage defects by judging the relationship between the horizontal distance D of adjacent nodes and the distance S between adjacent gold fingers.The relationship between the distance D in the direction of the adjacent nodes and the distance S between adjacent gold fingers is used to achieve the detection of short-circuit and breakage defects of FPC gold fingers.The feasibility of the FPC defect detection method based on the skeleton extraction is demonstrated through experiments conducted on the testbed.3)For FPC gold finger scratch,contamination,foreign matter and leakage plating defects,in order to solve the problems of slow speed and non-rotation invariance of traditional stencil matching,the research of fast stencil matching method based on circle projection and radial projection is carried out.The matching speed is accelerated by using Gaussian pyramid downsampling,jumping step and coarse-then-refine matching strategy,matching the position by circular projection,matching the angle by radial projection,differencing the matched image with the standard image,removing some pseudo-defects by mathematical morphology,and combining the distribution of defects in vertical direction,frequency,amplitude and color to achieve the detection of FPC scratches,contamination,foreign matter and The detection of defects in FPC is achieved by combining the distribution of defects in vertical direction,frequency,amplitude and color.The feasibility of the FPC defect detection method based on template matching is demonstrated by conducting experiments on a testbed.