머신 비전을 이용한 자동차 부품 검사 시스템에 관한 연구

Abstract

Machine vision system has widely been used for semiconductor inspection, PCB inspection and so on. Despite the rapid development of the automobile industry, there are few inspection systems for automobile parts. This study suggests machine vision system that was developed to automatically inspect automobile parts. The gear box has various models and is assembled from bolts, clips, fixed pads and so on. The gear box inspection system is suggested to inspect product assembly state. The inspection algorithm for the gear box allows Image binarization for coved structure of plastic products, image noise removal using morphology operation, labeling method for candidate regions, image filtering and calibration method for defect inspection. Also, the side protector inspection system is suggested to inspect surface defects. The inspection algorithm for the side protector is processed largely by a method using average of pixels in region of interest and other methods dividing the area and checking the presence of the object. To verify the effectiveness of these systems, the accuracy evaluation and the processing time evaluation was conducted.