Research On The Key Technology Of Electroplating Plug-in Appearance Quality Detection Based On Machine Vision

With the development of the electronics industry,in order to enrich the functions of electronic products,various electroplating plug-ins with different functions are indispensable as the basic parts.The defect quality inspection technology for electroplating plug-ins in most factories is still in the stage of manual sampling or semi-automatic detection,which is difficult to meet the mass quality inspection task.At the same time,manual sampling or semi-automatic detection requires more human resources and causes inaccurate test results.Therefore,in order to solve this problem,based on machine vision image processing technology,this paper designs a detection scheme for detecting the surface appearance defects of plug-ins and realizes the automatic detection of defects.Firstly,according to the research status of workpiece quality detection at home and abroad,aiming at the limitations of contact measurement and intermediate media detection methods,in order to effectively automate the appearance quality detection of electroplating plug-ins,this paper chooses the detection method based on machine vision and image processing,and studies the defect detection algorithm.Secondly,in the stage of image preprocessing,this paper analyzes the commonly used threshold segmentation and image denoising algorithms,proposes an ROI extraction algorithm based on Otsu global threshold segmentation combined with morphology.Through the performance test of the algorithm,the effect of separating the surface part of the plug-in image from the background is better.The image denoising correlation algorithm is studied and implemented.After experimental comparison,the gaussian filter algorithm with the best smoothing effect is finally selected.Then,aiming at the characteristics of the abrasion defects on the plug-in surface,this paper studies the algorithm of detecting the abrasion defects with threshold segmentation based on the small difference between the gray scale of the abrasion defects and the surrounding area,and implements the improved local dynamic threshold segmentation algorithm based on the global threshold.Aiming at the directional characteristics of scratch defects,a scratch extraction algorithm based on Gabor filtering was proposed.The effectiveness and superiority of the proposed algorithm in the task of flaw detection and extraction were proved by experiments.Moreover,the detected scrapes and scratch defects were classified and statistically analyzed by support vector machine.Finally,defect detection system development,choose the right camera image acquisition platform,and will write the Halcon software platform defect detection algorithm ported to Microsoft Visual Studio2012 environment,develop the electroplating plugin defect detection system based on machine vision,treatment of electroplating plug-in defect detection and classification experiment,the detection system is verified by the experimental results using detection algorithm proposed by this paper for rapid detection of defects in the image and classification,to achieve the requirement of the test.