摘要
传统生产工艺中,涂胶缺陷通过人工进行检测,其效率低、准确率差,且受人员责任心及主观经验影响。近年来,越来越多的汽车企业采用机器视觉技术进行缺陷检测。针对白车身生产过程中的某一具体涂胶工艺展开了探讨,利用2D面阵相机,通过图像分割、目标提取、骨骼化、B样条曲线拟合等手段,对涂胶区域及路径进行了分析和拟合,进而完成了对涂胶宽度、长度、位置、多涂、少涂、断胶等缺陷的检测工作。此外还对视觉系统进行了相机标定和手眼标定,对于涂胶断胶、少涂等缺陷能够引导机械臂自动进行补涂,自动修复相应的涂胶缺陷。
In traditional production processes,adhesive defects are manually inspected,which is inefficient,inaccurate,and influenced by human responsibility and subjective experience.In recent years,an increasing number of automotive companies have adopted machine vision technology for defect detection.This paper discusses a specific adhesive coating process in the white body production process.By using a 2D array camera and employing methods such as image segmentation,object extraction,skeletonization,and B-spline curve fitting,the adhesive area and path are analyzed and fitted.Additionally,camera calibration and hand-eye calibration are performed on the vision system.This allows the automatic re-application and repair of adhesive defects,such as adhesive breakage and under-application,guided by a robotic arm.
出处
《工业控制计算机》
2024年第7期21-22,25,共3页
Industrial Control Computer
关键词
机器视觉
白车身
涂胶
缺陷检测
自动修复
machine vision
BIW
adhesive coating
defects inspectation
automatic repair
