Research On Application Of Automotive Parts Production Line In Machine Vision

In recent years, with the development of the automobile industry, China’s auto partsindustry has been developing rapidly. As an indispensable part of vehicle, auto partsproduction must be followed by the rhythm of vehicle production, thus causing theproduction of auto parts faced with the dual challenges of efficiency and quality.Traditional manual detection methods not only cost a lot of labor, but also have a lowdetection efficiency and quality. Therefore, the study of auto parts production lineautomatic detection system with high efficiency and quality has become an urgent demandfor the domestic auto parts enterprises. Machine vision detection is a technology relying oncomputer science and insteading the manule detecion of machine, applying it to auto partsdefect detection can not only realize the automatic detection, reduce labor cost, and cangreatly improve the efficiency and quality of detection.Therefore, research of auto partsdetection technology based on machine vision has gradually become an important researchdirection of domestic universities and enterprises.This paper takes S company auto parts production line as the research object, anddiscussed the key technology of auto parts detecion based on machine vision, the mainwork done is as follows:(1) Through the investigation of S enterprise auto parts production line, manyproblems have been discovered from the current detection method such as low detectionquality of production color, location and defect. According to the survey results, the paperdetermined the target needs of machine vision detection system.(2) Based on the analysis of the selection of machine vision equipment in industrialdetection application and combined with high precision of product, this paper sets suitableenvironmental factors for machine vision detection system such as cameras, lens, computer,lighting etc. The results of the image processing showed that selected equipment andenvironmental factors meet the accuracy requirements of detection system.(3) Based on the analysis of the two-step method and Zhang Zhengyou cameracalibration method, the paper selects the Zhang Zhengyou multi-plane templates camera calibration method for CCD camera, providing guarantee for accurate measurement of thedefect size.(4) Taking S company seat products as an example, with target demand of detectionsystem, this paper studiesthe detection methods on product color, location and defects. Byimage preprocessing, image enhancement, image segmentation, mathematical morphologyprocessing and other means, the color, location, defect and defect size measurement.of seatproducts are finally realized.