Robotic Arm Item Detection And Grabbing Based On ROS Platform

With the development of the times,the application of industrial robotic arms has become more and more widespread today.However,the control methods of online teaching and offline programming are still the mainstream of manipulator control nowadays,which also leads to some defects in portability and autonomy of industrial manipulators,which can only be carried out in a fixed environment and fixed process scenarios.Work,and between different robot arms,the control methods are usually not transplanted to each other.Therefore,how to improve the transplantability and autonomy of the robotic arm in actual production has become a research hotspot in recent years.Aiming at the above two points,this paper proposes a robotic arm visual contour detection and grasping system based on the ROS platform.Using machine vision and ROS system to improve the autonomy and portability of the robotic arm,the main work content is as follows:(1)This article first,for the experimental goal,divides the overall system into three parts,namely the vision part,the robotic arm part and the system integration part.(2)In the visual part,the Realsense D435i camera was calibrated with the help of Zhang Zhengyou's calibration method to obtain the internal parameters of the camera.Subsequently,in order to obtain the target contour and center point in the original image,the storage form of the image in the computer is first explained,and methods such as color space conversion,smoothing,edge detection,contour extraction,image extraction,and template matching are used to perform the image processing.Process,and use the contour Hu moment to improve the matching method,and finally draw the outline of the target object and mark the center point.(3)In the manipulator part,in order to control the manipulator clearly and effectively in ROS,the description of the coordinate space and its transformation method involved are explained,and then the DH parameter is used to represent each link in the manipulator.The position relationship between the robot arm is analyzed,and the forward and inverse kinematics analysis is carried out for the manipulator,and then the fifth-order polynomial interpolation method is determined as the trajectory planning method of the manipulator.After that,in order to establish a visualized robot arm model,in the ROS environment,the xacro file was selected as the description file of the robot arm,and the robot arm was configured related,and finally the robot arm model was successfully established and implemented in the Rviz interface.display.(4)In the integration part,CvBridge is used to effectively integrate the image processing results with the ROS system,and the nine-point calibration method is used to establish the coordinate relationship between the camera and the robotic arm.Finally,the grasping experiment is carried out for different targets.The result is an error analysis to verify the feasibility of the scheme.