摘要
为实现对复杂工件的快速在线测量,构建了基于结构光的机器人视觉测量系统,提出一种该系统中结构光测头内参数和手眼关系同时标定的方法。标定时,在机器人工作空间内固定一个平面靶标作为参考物,精确控制机器人末端带动结构光测头做平移运动和变位姿运动,获取不同位姿下的靶标图像。通过平移运动,同时标定结构光测头内参数和手眼关系中的旋转矩阵;通过变位姿运动,同时标定摄像机内参数和手眼关系中的平移向量。考虑到旋转矩阵和平移向量分离标定会存在误差累积,采用非线性优化方法对手眼关系做进一步优化。对标准球及工件进行扫描测量实验,实验结果表明该标定方法简单高效,能保证测量系统具有较高的测量精度,适用于工业现场。
In order to realize rapid online measurement of complex workpiece, a robotic flexible optical measurement system based on structured light is established. A simultaneous calibration method for the intrinsic parameters of struc- tured light sensor and hand-eye relation of robot is proposed. In the calibration process, a planar target is fixed as ref- erence substance in the robot workspace. The robot end-effector is accurately controlled to perform translational mo- tions and pose-changed motions, and the camera captures the image of the planar target in different poses. The intrinsic parameters of the structured light sensor and hand-eye rotation matrix are calibrated simultaneously by the translational motions. The camera intrinsic parameters and the hand-eye translation vector are calibrated simultaneously by the pose-changed motions. Considering the error propagation between the orientation and position calibration, the hand-eye relationship is further optimized. Scanning experiments are carried out for a standard sphere and a complex work- piece. The results show that the calibration method is simple, efficient and has high accuracy, and it is well suitable for the calibration work of on-line measurement system in industrial scenes.
出处
《激光与红外》
CAS
CSCD
北大核心
2017年第9期1142-1148,共7页
Laser & Infrared
基金
国家自然科学基金项目(No.61571408)资助
关键词
结构光测头
内参数
手眼关系
标定
机器人
structured-light sensor
intrinsic parameters
hand-eye relation
calibration
robot
