基于单目视觉的工业机器人智能抓取研究

摘要

With reference to the automatic assembly problems ocurred in industrial assembly line ,a monocular vision-based robot with automatic identification and intelligent grasp is designed.Matching relationship matrix is established by applying rules of maximum degree of membership.And decision matrix is designed as well in order to recognize the workpieces fuzzily. Meanwhile ,the method of second moment of inertia is applied to determine the axia.The experiment which allows the robot to recongnize and grasp the workpiece on the assembly platform shows that the better recognition effect and real-time is achieved and thus provide necessary information for industrial robot to grasp the target object intelligently.%针对工业装配生产线的自动装配问题,设计了一种基于单目视觉技术的机器人自动识别与智能抓取的结构.运用最大隶属度规则建立匹配关系矩阵,构建决策矩阵达到工件的模糊自主识别目的,并采用基于二阶惯性矩的轴向确定方法.在机器人装配作业平台上进行了工件识别与抓取的实验研究,实验结果表明,该方法能够获得很好的识别效果和实时性,为工业机器人智能抓取目标提供了必要的信息.