基于双目视觉的零件位姿测量系统研究

Research on Part Pose Measurement System Based on Binocular Vision

在机器人自动加工系统中,针对待加工零件存在体积大、异形等导致工件难以精确定位的问题,提出了一种基于双目视觉的零件位姿测量系统。在该位姿测量系统中,通过在待加工零件上放置不等腰直角三角形参照物的方法来对其进行局部双目拍摄。对拍摄后的图像采用Harris角点特征检测参照物角点并匹配,进而求解参照物的角点世界坐标,然后通过坐标变换完成对待加工零件的初始位姿测量。并建立了位姿测量系统的数学模型,通过MATLAB语言编写了位姿求解计算程序。最后搭建了位姿测量系统实验平台,联合C#语言设计了一款零件位姿测量系统软件,并进行了位姿测量实验。本视觉系统相对于市场上机器人加工系统中的视觉系统成本很低。实验结果表明该系统能够有效求解待加工零件的位姿,为修正机器人加工系统中的加工轨迹提供了技术支持。

In the robot automatic processing system,in order to deal with the problems of large volume and abnormal shape of the parts to be processed,it is difficult to accurately locate the parts.A part pose measurement system based on binocular vision is pro⁃posed in this paper.In this pose measurement system,a partial binocular shot is performed by placing a non-isosceles right triangle reference object on the part to be processed.For the captured image,Harris corner features are used to detect and match the corner points of the reference object,and then the world coordinates of the corner points of the reference object are solved.And then the ini⁃tial pose measurement of the part to be processed is completed through coordinate transformation.The mathematical model of the pose measurement system is also established,and the calculation program for the pose calculation is written through MATLAB language.Finally,an experimental platform for the pose measurement system is built,and a part pose measurement system soft⁃ware is designed in conjunction with C#language.And a pose measurement experiment is carried out.Compared with the vision system in the robot processing system on the market,the cost of this vision system is very low.The experimental results show that the system can effectively solve the pose of the parts to be processed,which provides technical support for correcting the processing route in the robot processing system.