Development Of FPC Solder Defects Detection System Based On Machine Vision

FPC(Flexible circuit board)belongs to PCB(printed circuit board),which is a key electronic interconnection device of electronic products.It has the characteristics of high reliability and excellent flexibility,and is widely used in modern electronic products.This paper conducts theoretical analysis and experimental research on the solder defects of specific types of FPC based on machine vision technology,and develops the FPC solder defect detection system.The Solder defects of FPC include offset defects,reveal-copper defects,short-circuit defects and tin bread defects.The main work of this paper is as follows.1.Finished the system hardware selection and layout.Considering the requirements,feasibility,economic,and lighting analysis of project,the hardware of the system,including cameras,lens and light source,is selected.According to the installation test,the system hardware layout is completed.2.Designs offset defect detection Method.Firstly,Image ROI(Region Of Interest)technology is applied to improve the speed of image processing algorithm hugely,and Laplacian sharpening algorithm is used for ROI image enhancement preprocessing.Secondly,the template matching algorithm is selected to design the detection method of offset defects.The principles of general template matching algorithms,pyramid image search strategy and multi-angle template strategy in the template matching algorithm are introduced.The NCC(Normalized Cross Correlation)template matching algorithm of Halcon is selected for the design of offset detection method.Finally,the offset defect detection is designed by the position relation of each template which is calculated by template matching algorithm,and the test results of the method is analyzed.The test result shows that the precision of the offset defect detection is over 99.9%,which meets the requirements of the project.3.Designs the solder defect detection method.First of all,the detection methods of reveal-copper defect,short-circuit defect and tin bead defect are put forward.The detection method of reveal-copper defect is to extract the copper region by using the methods of RGB color channel separation,OSTU threshold segmentation,morphologyclosing operation and contour area screening,and its’ detection result depends on the intersection calculation results of copper region and solder template region.The short-circuit defect detection method uses the methods of RGB color channel separation,OSTU threshold segmentation,morphology closing operation and contour area screening to extract the solder area,and its’ detection result depends on the intersection calculation results of solder area and solder template area.The tin bread defect detection method uses OSTU threshold segmentation,morphology closing operation,Canny operator contour edge extraction and contour center calculation to calculate the location information of solder contour area,and its’ detection result depends on the coordinates of solder contour area.In addition,the test results of the three kinds of solder defect detection methods are analyzed,and the test results shows that comprehensive precision of the solder defects detection methods is over 99.5%,which meets the test requirements of the project.4.Designed software systems.Upon the theoretical research,the FPC solder defect detection system based on machine vision technology is developed,which meets the actual testing requirements.