摘要
将双目立体视觉系统引入到激光修复再制造机器人中,可以实现机器人对修复物体缺陷的自动识别,给出最佳修复路径和工艺参数。标定问题是视觉系统基本且关键的工作。将两个摄像机安装在机器人末端,随机器人一起移动,对二维平面靶标进行标定实验。通过机器人示教盒改变摄像机的位置,用左右摄像机各采集不同位置的靶标图片,采用OpenCV编程,求得摄像机的内外部参数,再用Matlab完成了机器人手眼标定。借助机器人示教盒显示的数据,检测了摄像机标定精度。结果表明,最大相对误差1.08%,满足激光修复再制造要求。
Binocular stereo vision system is equipped in laser remanufacturing robot.The robot can automatically find the defect of objects to be repaired and give the best repairing path and technological parameters.Calibration is the basic and key task of vision system.Two cameras are installed in the robot terminal and calibrate the two-dimentional(2D) plane target.The right and left cameras acquire target images in different positions by changing the camera′s positions through the robot teaching pendant.Internal and external camera parameters are calculated by OpenCV programming and robot hand-eye calibration is completed by Matlab.Calibration accuracy is analyzed with the help of robot teaching pendant.The result shows that the maximum relative error is 1.08%,which meets the requirements of laser-repairing remanufacturing.
出处
《中国激光》
EI
CAS
CSCD
北大核心
2010年第7期1868-1872,共5页
Chinese Journal of Lasers
基金
国家自然科学基金(60478004)
天津市科技支撑计划重点项目(08ZCKFGX02300)资助课题
关键词
激光技术
再制造
机器人
机器视觉
标定
laser technique
remanufacturing
robot
machine vision
calibration