Development Of Online Welding Quality Inspection System Based On Machine Vision

In recent years,the electronics manufacturing industry has made great strides,and the output of printed circuit boards as carriers has soared.The quality inspection of its solder joints has attracted much attention.Traditional methods generally rely on manual labor,have low accuracy,and take a long time.At present,they cannot meet the demands of automated,high-speed,high-precision production,and they need to be improved.This paper researches and develops a solder quality inspection system based on machine vision technology.The system uses a motion platform to drive the Keyence profiler for detection.Compared with the traditional method,the equipment has the strong point of rapid and high accuracy.The system extracts feature from a three-dimensional point cloud of the welding point effectively reduces the complexity of the processing of the welding point.In this paper,based on the research on the technical theories and key technologies of solder quality inspection,combined with the actual inspection requirements,an online solder inspection method based on machine vision is used.According to the actual characteristics of the welds to be inspected,a line structured light camera was selected as the point cloud acquisition device from various 3D cameras,and a detection method combining point cloud processing and image processing was proposed.By using the Open CV,combined with image processing theories such as point cloud filtering,spatial transformation,point cloud segmentation,image threshold,image morphology operation,feature extraction,etc.,the point cloud of the solder to be inspected is analyzed and processed to obtain the Width,height,surface area,volume and other information.Among them,according to the actual appearance of the solder,a method based on the reference surface is used to achieve high-precision solder separation.Based on four types of solder defects: more solder,less solder,solder adhesion,and missing solder,the extracted solder features are classified using the K-nearest neighbor algorithm.The results of solder defect classification are analyzed to determine whether the solder quality is acceptable and statistical analysis of the undetected ratio,error detection ratio,and accurate detection ratio of solder inspection.The overall scheme of online solder quality inspection equipment is designed,and the hardware selection is discussed in detail in combination with practical application experience,including how to select important components in machine vision inspection systems such as line-structured light cameras,motion control cards,and industrial control machines.Set up an online solder defect detection software platform,provide a good human-computer interaction interface,solve the problems in the actual debugging,ensure that the solder detection system runs quickly and accurately,realize the real-time online detection of the defect detection system,and test software the overall workflow is described.The online solder quality three-dimensional inspection system greatly reduces the outflow of solder defective products and has a more practical reference value in the industry.