通过一台单目相机采集目标物体图像,采用灰度化、中值滤波处理、Canny边缘检测、直线检测实现对图像的处理,识别定位目标空间信息.根据机械臂模型的建立和对正、逆运动学的求解,结合ROS和OpenCV实现对目标的识别和定位功能,完成对机械臂关节转动角度的求解,实现了机械臂关节变量空间和笛卡尔坐标空间双可控.针对当前带电紧固螺丝作业机器人存在的弊端,设计将人工智能启发式算法与机械臂笛卡尔空间路径规划设计结合,由算法自动规划得出最短路径,实现机械臂末端跟踪控制,并确保环境重建和任务设计不超界,进而实现自主作业.最后进行了机械臂的抓取全过程仿真和实验.结果表明:机械臂在抓取过程中运行平稳,能够准确识别目标并进行抓取.
The target object image is collected using a monocular camera,and processed using grayscale,median filtering,Canny edge detection,and line detection to identify and locate the spatial information of the target.Ac-cording to the establishment of the robot arm model and the solution of the forward and inverse kinematics,com-bined with ROS and OpenCV to realize the target recognition and positioning function,complete the solution of the rotation angle of the robot arm joint,so that the joint variable space and Cartesian coordinate space of the manipula-tor can be controlled and can be evolved mutually.To avoid the disadvantages of the current live screw tightening robot,the design combines the artificial intelligence heuristic algorithm with the Cartesian space path planning and the design of the robot arm,and the algorithm automatically plans the shortest path to realize the tracking control of the end of the robot arm,and ensures the environment reconstruction and task design not exceeding the boundary,so as to realize the autonomous operation.Finally,the whole grasping process of the robot arm is simulated and tested.The results show that the robot arm runs smoothly in the grasping process,and can accurately identify the target and the grasp.