双目视觉引导机器人定位抓取技术的研究

目前大部分工业装配生产线中机器人抓取待装件都是预先将固定类型的工件进行机械定位。为了提高机器人柔性抓取能力,实现了双目视觉引导机器人的抓取系统。文章通过定位工装板来间接定位工件的方法来完成双目视觉引导机器人抓取,首先通过双目视觉系统拍摄的图像识别工装板上6个定位孔,计算孔中心在相机坐标系下的位置,利用这6个定位孔中心空间位置用最小二乘法拟合出最优的工装板平面,由于工件相对于工装板的位置固定,可求得工件相对于相机坐标系的位置,进而通过标定参数转化为在机器人基坐标系下位置和姿态,供机器人抓取工件。经过多次工件抓取实验,分析数据得出,系统距离定位误差小于0.5mm,满足工业生产要求且适应于多种类型工件的抓取,具有通用性。

Research on Robot Positioning and Grasping Technology Based on Binocular Vision

At present,most robots in the industrial assembly line have to pre-position the fixed type of workpiece for mechanical positioning.In order to improve the flexible grasping ability of the robot,the robot grasping system based on binocular vision is realized.In this paper,the system are captured by locating the tooling plates to locate the workpiece indirectly.First,the images captured by the binocular vision system are used to identify the six positioning holes on the tooling plates to calculate the position of the holes center under the camera coordinate system.Using the least square method to fit the optimal plate plane by using the six location holes in the center space.Because the position of the workpiece relative to the plate is fixed,the position of the workpiece relative to the camera coordinate system can be obtained and transformed by the calibration parameters for the base coordinate system position and posture for the robot to grab the workpiece.After several workpiece grasping experiments,the analysis of the data shows that the system positioning error is less than 0.5mm,to meet the requirements of industrial production and adapt to a variety of types of workpiece,with versatility.