Research On The Detection Method And System Design Of Assembly Defects In Polar Electronic Components

With the rapid development of new generation information technology and its deep integration with manufacturing,intelligent manufacturing has entered a new stage from point to surface.Printed Circuit Board(PCB),as the most basic component in electronic devices,is an important cornerstone for the rapid development of intelligent manufacturing.Component defect detection is a key and necessary link in the Printed Circuit Board Assembly(PCBA)process,especially for polarized components with assembly direction requirements.Post-detection can greatly reduce the risk of property damage and harm to life caused by component short circuits.Due to the rapid development of the industry,traditional manual visual inspection(MVI)and in-circuit testing(ICT)schemes are no longer able to meet the current market demand.Automated Optical Inspection(AOI)schemes have gradually become the main solution in the field of PCB defect detection due to their high performance and non-contact advantages.This article focuses on the detection of polarized component assembly defects on plug-in PCBs and designs a fast and high-performance detection system based on a visual scheme.The main work completed is as follows:(1)Design and build a polarized component assembly defect detection experimental platform.According to the application requirements of the production line,the overall framework of the system is established.The key equipment for image acquisition is selected.Through preliminary research and illumination experiments,the light source type of the system and the structure of the equipment box are determined.(2)A polarized capacitor detection algorithm based on polar coordinate transformation and Support Vector Machine(SVM)fusion is proposed to meet the detection requirements of polarized component assembly defects,with the electrolytic capacitor as a typical circular surface element.Two detection schemes are developed based on traditional image algorithms and machine learning,respectively.Experimental results show that each type of scheme has advantages in detection accuracy and time consumption.To balance accuracy and time consumption,a polar coordinate transformation and SVM fusion-based capacitor polarity detection algorithm is proposed.(3)A 2-norm symmetric region singularity degree comparison method is proposed for rectangular surface elements such as sockets and diodes to determine polarity.The algorithm compares the singularity degree of the feature vectors of the tested element image and the template image in the symmetric region.Experimental results show that this algorithm exhibits strong universality for various sockets,even diodes and other different types of rectangular surface elements with different shapes,colors,and sizes,and has high detection accuracy.(4)A software operation platform is developed and field-tested in the PCBA workshop.According to actual needs,a concise operation interface and highly integrated functional module are designed to facilitate frontline workers.In the testing phase,eight batches of PCBs in the PCBA production line were randomly sampled,with a total of 5,360 PCB boards and 198,110 components detected.Experimental data show that the average accuracy of PCB matching,component positioning,and polarity detection are 99.85%,98.96%,and 99.64%,respectively,with an overall accuracy of98.46% and a false negative rate of 0.03%.