基于视觉的机械臂目标抓取研究

Research on Target Grasping for Manipulators Based on Vision

工业机器人通常在工作位置已知的情况下工作,一旦位置发生改变便无法继续工作,为适应这种改变,提出将视觉传感器与工业机器人相结合。建立图像与三维空间的转换关系,对相机光路进行分析建立相机模型,并对镜头畸变引起的误差进行修正;通过相机获得物体的图像数据,对图像数据进行处理获得二值图像;对物体在图像中的质心位置进行计算,分析质心位置偏差产生的原因并进行修正,完成对物体的定位。最后对物体进行了定位精度测试和抓取实验。

Industrial robots usually work in a known position and under a known condition,in the case of which,they would not able to normally working once the position has been changed.In order to adapt to the change,a method that combines vision sensors and industrial robots was proposed in this article.Establishing the converting relationship between images and three-dimensional（3 D）space,the camera imaging path was analyzed and the model of camera was built,hence correcting the errors due to lens distortion.Furthermore,the image data of objects caught by cameras would be processed in order to obtain binary images.Then the centroid positions of the objects in the images would be calculated in order to analyze the reason of position deviation and correct it,so that the object positioning would be completed.Finally,the positioning accuracy test and grasping test would be carried out by experiment.