基于双目视觉的物体识别定位与抓取

为解决目标识别精度低、定位与抓取配合困难问题,通过对算法的改进研究了物体的识别定位与抓取,包括识别目标物体、对于目标物体形心的位置三维重建、对于目标物体的姿态估计、对于手眼标定后的抓取4个方面。在目标识别中,首先通过快速鲁棒特征(speeded up robust features,SURF)算法结合Grabcut算法提取出目标。在位置求解中,利用模板匹配求取目标物体形心的世界坐标。在姿态估计中,算法的流程为:利用匹配点对求取左图中物体母线斜率,再随机取斜率等于左图母线斜率的两点,通过两点的世界坐标求出目标物体的姿态。在抓取中,采用的眼在手上,先建立工件坐标系,再进行坐标转换,通过机器人参数求得逆解。研究结果表明:误差均在较小范围内且机器人可在有效工作范围实现抓取。可见算法可靠以及整体实验的正确性。

Object Recognition and Capture Based on Binocular Vision

In order to solve the problem of low target recognition accuracy, it is difficult to locate and grasp. Through the improvement of the algorithm, the object is identified and captured. It is divided into four aspects: the first recognition target object; the second three-dimensional reconstruction of the position of the target object centroid; the third attitude estimation for the target object; and the fourth for the hand-eye calibration. In the target recognition, the target is first extracted by the SURF algorithm combined with the Grabcut algorithm. In the position solving, the world coordinates of the centroid of the target object are obtained by template matching. In the attitude estimation, the flow of the algorithm is to use the matching point pair to obtain the slope of the object bus in the left picture, and then randomly take two points whose slope is equal to the slope of the left picture bus, and obtain the attitude of the target object through the world coordinates of the two points. In the grab, the eye used in the hand, first establish the workpiece coordinate system, then coordinate transformation, and obtain the inverse solution through the robot parameters. The results show that the error is in a small range and the robot can achieve the grab in the effective working range. The reliability of the algorithm and the correctness of the overall experiment can be seen.