Abstract: This paper presents a framework of tracking and detection system based on stereo vision for 6-DOF series robot. It is composed of image capture, camera calibration, tracking based on CamShift algorithm, Speeded Up Robust Features(SURF) algorithm and reconstruction of points, position and orientation measurement. This framework can be used for real-time tracking of the robot by CamShift algorithm, search and outline the robotic arm operator position in the current area. For the robotic arm to be tracked, it should take SURF algorithm for feature point detection and stereo matching. The above method is used for robot position and orientation detection experiment. Experimental results show that the combination of the two algorithms is especially fit for the 6-DOF robot tracking and detection and complex motion in space.

Key words: series robot, tracking detection, position and orientation measurement, stereo matching, CamShift algorithm, Speeded Up Robust Features(SURF) algorithm

摘要： 建立一种六自由度串联机器人视觉跟踪检测系统框架，包括图像采集、摄像机标定、机器臂跟踪检测、机器臂位姿建模与计算等。提出利用CamShift算法对机器人进行在线粗跟踪，搜寻和画定出机器臂操作器在当前窗口的区域位置。对跟踪到的机器臂按照SURF算法进行特征提取与立体匹配。该方法被用于对串联机器人位姿检测进行实验。实验结果表明，2种算法的结合适用于六自由度串联机器人在空间复杂运动的跟踪检测。

关键词: 串联机器人, 跟踪检测, 位姿测量, 立体匹配, CamShift算法, SURF算法

CLC Number: 

• TP18