基于机器视觉的直角机器人直线度检测方法研究  

Research on Straightness Detection Method of Cartesian Robot Based on Machine Vision

作　　者：曹继项 张军[1] 陈国栋 CAO Jixiang;ZHANG Jun;CHEN Guodong(School of Construction Machinery,Chan g’a n University,Xi’an Shaanxi 710064,China;School of Mechanical and Electric Engineering,Soochow University,Suzhou Jiangsu 215000,China)

机构地区：[1]长安大学工程机械学院,陕西西安710064 [2]苏州大学机电工程学院,江苏苏州215000

出　　处：《机床与液压》2018年第22期108-111,145,共5页Machine Tool & Hydraulics

摘　　要：针对现有设备测量和标定机器人末端直线度费时、成本高的问题,提出基于机器视觉的非接触快速检测方法。采取直接采集机器人末端圆靶标的圆心位置信息间接提取机器人末端的位置信息,采用改进的Canny边缘检测对目标边缘粗定位,用改进的随机霍夫圆变换进行靶标圆心位置信息的快速精确提取。用最小二乘法对末端的直线度做出评定。通过修改控制直角坐标机器人的代码对直线度误差做出补偿。实验结果表明:经过两次补偿后,代码控制的机器人末端相应直线的直线度精度可达约0. 1 mm。Aiming at the problem of time-consuming and high cost of measuring and calibrating the robot at the end of the existing equipment,a non-contact rapid detection method based on machine vision was proposed.The position information of the robot end was extracted by using the center information of the robot head end circle,and the improved Canny edge detection was used to locate the target edge.The improved Hough circle transform was used to extract the target position information quickly and accurately.The straightness of the ends was evaluated by the least squares method.The straightness error was corrected by modifying the code that controls the cartesian robot.The experimental results show that after two compensation,the straightness accuracy of the straight line at the end of the robot controlled by code is up to 0.1 mm.

关 键 词：机器视觉 边缘检测 直线度 随机霍夫变换 最小二乘法 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]