机器视觉主导的机械臂动态抓取策略研究  

Research on Dynamic Grasping Strategy of Robot Arm Dominated by Machine Vision

作　　者：王校峰 王建文[1] 曹鹏勇 杨云茂 WANG Xiaofeng;WANG Jianwen;CAO Pengyong;YANG Yunmao(School of Mechanical and Power Engineering,East China University of Science and Technology,Shanghai 200237,China)

机构地区：[1]华东理工大学机械与动力工程学院,上海200237

出　　处：《机床与液压》2022年第17期38-42,共5页Machine Tool & Hydraulics

摘　　要：针对传统工业机器人在移动目标的抓取中存在灵活性和位姿适应能力差的问题,以开源机器人操作系统(ROS)为平台进行基于视觉反馈的动态抓取策略研究。在ROS中搭建UR5机械臂动态抓取系统,并对机械臂进行正逆运动学分析;提出基于RGB颜色空间和SITF算法结合的新型特征识别方法以提高识别效率;设计注视逼近的视觉反馈控制策略,实现机械臂对传送带上移动目标的抓取操作。结果表明:所提策略识别成功率达到99.6%,逼近抓取成功率达到98.33%,验证了机器视觉主导的反馈抓取算法基本满足工程实际应用,有较高的可行性和准确性。Aiming at the problem of poor flexibility and positional adaptability in the cracking of traditional industrial robots in the movement target,a dynamic grasping strategy based on visual feedback was performed in an open source robot operating system(ROS).The UR5 manipulator dynamic grasping system was constructed in ROS,and the forward and inverse kinematics of the manipulator were analyzed;based on RGB color space and SITF algorithm,a new feature recognition method was proposed to improve the recognition efficiency;a visual feedback control strategy was designed for gaze approaching to realize the grasping operation of the moving target on the conveyor belt by using the mechanical arm.The results show that by using the designed strategy,the recognition success rate can reaches 99.6%,and the approximate grasping success rate can reaches 98.33%,which verify that the feedback capture algorithm dominated by machine vision basically meets the practical application of engineering,and has high feasibility and accuracy.

关 键 词：工业机器人 机器视觉 动态抓取 目标识别 操作系统 

分 类 号：TP242.2[自动化与计算机技术—检测技术与自动化装置]