基于机器视觉的车身定位检测系统设计

Design of Vehicle Body Positioning-detection System Based on Machine Vision

某主机厂车身从滑板转接到自动运行吊具后,吊具定位销是否入车身孔的检测,采用传统的光电式传感器进行检测。然而,光电式传感器处理信息量小、功能单一,导致特定场景下误检的问题时有发生。针对该问题,设计了一种由PLC、照明、工业相机、镜头等构成的基于机器视觉的车身定位检测系统。由PLC触发相机采集车身特定部分的二维数字图像,通过图像处理获取车身的边缘特征,采用边缘检测和直线检测算法,实现吊具定位销是否入车身孔的检测。阐述了图像中灰度值变换确定边缘的原理,研究设计了系统方案,对系统硬件进行选型及对系统软件进行设计。与传统的光电传感器检测相比,采用机器视觉的检测系统能大幅提高检测的精准度。基于机器视觉的车身定位检测系统投产后的数据表明:该系统对吊具定位销是否入车身孔的检测准确率达到100%。

A certain automobile OEM to detect whether the positioning pin of the hanger enters the body hole after the body is transferred from the skillet to the hanger uses a traditional photoelectric sensor.However,the natural deficiency of photoelectric sensors often leads to the occurrence of false positives in specific scenarios.Vehicle body positioning-detection system based on machine vision consisting of PLC,lighting,industrial camera,lens,etc.was designed to address this issue.The PLC triggers the camera to collect two-dimensional digital images of specific parts of the vehicle body,and obtains the edge features of the vehicle body through image processing.Edge detection and line detection algorithms are used to detect whether the positioning pins of the hanger have entered the vehicle body holes.Elaborated on the principle of determining edges through grayscale value transformation in images,studied and designed a system scheme,selected system hardware,and designed system software.Compared with traditional photoelectric sensor detection,the detection system using machine vision can significantly improve the accuracy of detection.After the machine vision system get into running,the data shows that the accuracy of the system in detecting whether the positioning pins of the hanger enters the vehicle body holes reaches 100%.