摘要
针对工业生产中的轮廓不规则工件难以定位的问题。以UR5机械臂为研究对象,设计了一套视觉定位系统。文中借助视觉技术定位工件,实现机械臂自动抓取目标工件。改进Canny边缘检测算法,然后对检测到的工件轮廓,利用Hough投影法寻找目标轮廓的主轴,建立目标工件的最小外接矩形。实现目标工件的定位并确定目标图像位置,最终实现机械臂抓取目标工件。通过抓取实验证明,文中的算法满足实际需求,具有实际工程应用价值。
Aiming at the problems of irregular workpiece contour and difficult positioning in industrial production,a set of visual positioning system is designed with the UR5 manipulator as the research object.This paper applies vision technology to locate the workpiece and realize the automatic grasping of the target workpiece by the robotic arm.The Canny edge detection algorithm is improved,and then the Hough projection method is used for the detected workpiece contour to find the main axis of the target contour,and the minimum circumscribed rectangle of the target workpiece is established.The positioning of the target workpiece is realized and the target image position is determined,and finally the robot arm grasps the target workpiece.The crawling experiment proves that the algorithm in this paper meets the actual needs and has practical engineering application value.
作者
董阳
于洪鹏
台立钢
张禹
DONG Yang;YU Hongpeng;TAI Ligang;ZHANG Yu(School of Mechanical Engineering,Shenyang University of Technology,Shenyang 110870,China;Shenyang Institute of Automation of Chinese Academy of Sciences,Shenyang 110017,China;Institutes for Robotics and Intelligent Manufacturing of Chinese Academy of Sciences,Shenyang 110169,China)
出处
《机械工程师》
2022年第12期21-26,共6页
Mechanical Engineer
关键词
机器视觉
机械臂
轮廓检测
外接矩形
machine vision
robotic arm
contour detection
circumscribed rectangle
