Multi-view coordinate system transformation based on robot
Multi-view coordinate system transformation based on robot
复制标题
DOI：
10.1007/s11801-015-5066-2
复制
发表时间：
2015-11
影响因子：
0.9
通讯作者：
Limei Song;Peng-qiang Wang;J. Xi;Qinghua Guo;Huan Tang;Jing Li;Xiao-jie Li;Teng-da Zhu
中科院分区：
文献类型：
--
作者：
Limei Song;Peng-qiang Wang;J. Xi;Qinghua Guo;Huan Tang;Jing Li;Xiao-jie Li;Teng-da Zhu
点云配准对于大物体测量非常重要。提出一种基于机器人的坐标系变换测量方法。首先，为了获得外参，机器人移动到三个不同的位置以捕获三个目标的图像。然后利用机器人的已知参数和外参计算相机与工具中心点（TCP）坐标系之间的变换矩阵X，最后通过变换矩阵X将多视点坐标系变换到机器人坐标系。在机器人的帮助下，多视点点云可以轻松转化为统一的坐标系。通过使用机器人，测量不需要任何标记。实验结果表明该方法是有效的。
The registration of point cloud is important for large object measurement. A measurement method for coordinate system transformation based on robot is proposed in this paper. Firstly, for obtaining extrinsic parameters, the robot moves to three different positions to capture the images of three targets. Then the transformation matrixXbetween camera and tool center point (TCP) coordinate systems can be calculated by using the known parameters of robot and the extrinsic parameters, and finally the multi-view coordinate system can be transformed into robot coordinate system by the transformation matrixX. With the help of robot, the multi-view point cloud can be easily transformed into a unified coordinate system. By using robot, the measurement doesn’t need any mark. Experimental results show that the method is effective.