基于移动视觉的工业机器人分拣系统应用研究  

Application research of industrial robot sorting system based on mobile vision

作　　者：彭辉辉 刘飞飞 代云勇 陈臻阳 陈文涛 PENG Huihui;LIU Feifei;DAI Yunyong;CHEN Zhenyang;CHEN Wentao(School of Mechanical and Electrical Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,China;School of Electrical Engineering and Automation,Jiangxi University of Science and Technology,Ganzhou 341000,China)

机构地区：[1]江西理工大学机电工程学院,江西赣州341100 [2]江西理工大学电气工程与自动化学院,江西赣州341100

出　　处：《现代电子技术》2020年第20期26-30,共5页Modern Electronics Technique

基　　金：国家自然科学基金资助:离子交换法钨冶炼过程建模与控制策略(61364014)。

摘　　要：针对机器视觉结合工业机器人进行自动分拣工作存在的物料识别和定位等问题,提出一种基于运动视觉的物料测距和定位方法。主要原理是通过安装在工业机器人末端的摄像头在移动过程中,对世界坐标系不同位置进行连续单点拍摄,进而获得类似双目视觉测距的必要参数,实现运动测距和定位的功能。针对现有的一种圆柱体工件自动分拣作业,利用视觉库(OpenCV for Python)处理图像数据,在机器人运行轨迹内进行逐点采样计算。实验结果表明,单目运动视觉测距和定位效果良好,平均定位误差小于4%,该测距方法在一定条件下能够达到精度要求,具有降低自动分拣系统升级成本的实际意义。In allusion to the material identification and positioning in the automatic sorting work of machine vision combined with industrial robot,a method of material ranging and location based on motion vision is proposed.The main principle is to continuous single point shooting at different positions in the world coordinate system during the movement of the camera installed at the end of the industrial robot,so as to acquire the necessary parameters that similar to binocular vision ranging and realize the function of motion ranging and positioning.In allusion to the existing automatic sorting operation of cylinder workpiece,the OpenCV for Python is used to process image data,and point⁃by⁃point sampling calculation is carried out within the robot′s trajectory.The experimental results show that the monocular motion vision ranging and positioning is good,and the average positioning error is less than 4%.The ranging method can meet the accuracy requirements under a certain conditions.It has practical significance to reduce the upgrade cost of the automatic sorting system.

关 键 词：工业机器人 分拣系统 移动视觉 运动测距 运动定位 测距实验 

分 类 号：TN830-34[电子电信—信息与通信工程]