Research And Development Of System For Appearance Inspection Of Electronic Components

The electronics industry is facing greater market demand and development space becauseinformation technology is developping fast and continuing innovation, and it also drives bythe strong development of the electronic components industy, and nowdays the electroniccomponents are becoming more and more intelligence, integration and miniaturization, whichis the difficult in detecting the defect of the electronic components, and manual inspection isslow and low efficiency, and is vulnerable to subjective factors such as workers’ experienceand physical condition. In nowdays, with the continuous development of the computertechnology, machine vision technology has been widely used in pharmaceutical, packaging,automotive and other industries, it can be able to achieve varieties functions such as testing,judgment, recognition, measurement, besides that, it has provided a guarantee for theautomatic detection of the appearance of the electronic components that the innovation anddevelopment of digital image processing technology, camera and light source and otherhardware and software equipments. As a result, it is very urgent and critical to the enterpriseto research and develop the system for appearance inspection of electronic components whichrelys on machine vision technology. This research of this topic has a broad meaning andapplication prospects.This paper, firstly, introduces the knowledge associated with machine vision technology,on this basis, bulids a detection system with a glass rotating platform which can achieve takepictures. secondly, we analysis and generalize dimensions and defect of resistance and copper,and we do image pre-processing analysis for resistance and copper images; besides that, wecompared a variety of edge detection algorithms and proposed a defect detection algorithm ofchip components based on a improved self-adaptive dual-threshold canny algorithm, andrelized the simultanenous detection of all regions, meanwhile, we judge whether thecomponents are good or not by calculating the geometric characteristics of each blob in thedetection region using the area segmentation algorithm. Finally, it concludes that thisalgorithm not only reduces the detection time, but also improves the detection accuracy byanalyzing the results of the tests, and this topic provides a good basis to further improve theperformance of the detecting equipment.