Stereo Image Processing And Matching Technology For Robotic Grabbing Operations For Scattered Parts

The welding of the outer casing of the conventional solid rocket motor casing is prepared by the worker manually selecting the part and pressing on the casing to complete the pre-welding positioning.The number of external parts on the casing is large,distributed in different parts of the casing,the labor intensity of the workers is high,the welding position deviation is large,and the safety is difficult to be secured.Aiming at the above problems,the paper proposes an image processing and matching technology based on binocular vision for auxiliary welding of robots.The main work includes:(1)The robot and binocular camera coordinate system that satisfies the welding of the outer casing of the shell is established.The spatial position transformation matrix of the camera coordinate system and the robot coordinate system is obtained according to the experiment,and the coordinate system relationship is calibrated using the ROS platform.(2)For the binocular vision detection problem of scattered external components,the filtering method of depth image is studied.The comparison analysis experiments of mean filtering,Gaussian filtering,median filtering and guided filtering are given.The depth image processing based on guided filtering is given.Aiming at the problem of data loss in the depth image of scattered objects,an algorithm based on joint bilateral filtering is proposed to repair the image hole.For the image data jitter of depth image,the Kalman filter method is selected.The experimental results show that the depth data value is effectively reduced.range.(3)Propose a point cloud data storage structure based on k-d Tree,and perform voxel reduction and processing on the point cloud.By using the RANSAC method to remove the background noise,the method based on the European clustering segmentation divides the part point cloud into a single part point cloud,which improves the system processing efficiency.In the paper,the principal component analysis method is used to perform rough matching with the CAD model point cloud,and then the model point cloud and the point cloud of the part matching cloud are matched by ICP.The experimental results show that the matching algorithm meets the design requirements.(4)Finally,the paper designs a robot-aided welding system based on binocular vision,and gives the system hardware and software composition and information exchange mode.Based on the ROS platform,the experiment of robots grabbing target parts is realized,which meets the basic requirements of the welding task of the auxiliary housing.