基于机器视觉的差速器卡簧装配防错检测研究  

Differential snap ring assembly error proofing detection research based on machine vision

作　　者：任永强[1] 杜瑞 李军[1] 

机构地区：[1]合肥工业大学机械工程学院,安徽合肥230009

出　　处：《制造技术与机床》2018年第1期30-34,共5页Manufacturing Technology & Machine Tool

基　　金：2015国家智能制造专项(工信厅联装函2015J 415号);2015安徽省科技攻关项目(1501021004)

摘　　要：机械零部件的装配质量检测是提高装配线生产效率的重要环节。针对汽车变速箱装配线中差速器总成的卡簧装配防错检测需求,选用CMOS工业相机和光源在装配线上搭建视觉检测平台。利用C#和Emgu CV计算机视觉库开发检测软件,结合机器视觉理论和算法,提出了以面积周长比值快速检测圆形轮廓的方法,通过检测卡簧的双耳圆孔轮廓并计算圆心距离判断卡簧是否漏装错装。最后验证该检测软件满足实时性和准确性要求,有效降低人工检测的误检率,提高汽车变速箱装配线的自动化水平。Mechanical components assembly quality detection is an important step to improve the production effi- ciency of assembly lines. According to the demand for snap ring Assembly Error Proofing Detection of differential assy on automobile transmission assembly lines, a vision detection platform is set up with the selection of CMOS industrial camera and illuminant. Utilizing C# and EmguCV computer vision library to develop detection software and combining with machine vision theory and algorithm, a fast circle contour detection method is proposed based on area-perimeter ratio. And the existence of wrong and missing as- sembly of snap ring is judged by the detection on the circle contours Of snap ring lugs and computing the distance of centers of circles. In the end, the detection software is verified in meeting the demands for real-time capability and accuracy, thereby reducing false detecting rate effectively and raising the auto- mation level of automobile transmission assembly lines.

关 键 词：机器视觉 EmguCV 卡簧装配 防错检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TH166[自动化与计算机技术—计算机科学与技术]