Research On Robot Guidance And Positioning Technology Based On Point Cloud Processing

In recent years,with the development of China's economy,the trend of "replacing people with machines" has become more and more obvious.More and more robots are used in industrial scenes.Robot guidance and positioning technology has become an important technology in the field of industrial automation.The traditional robot guidance and positioning technology is mainly based on 2D images and for neatly placed workpieces.It is difficult to solve the problem of random bin picking.Therefore,in this paper,the robot guidance and positioning technology based on point cloud processing is used to solve the problem of random bin picking.After completing the DH parameter calibration to ensure the accuracy of the position of the end effector of the robotic arm,and completing the hand-eye calibration to ensure the accuracy of the transformation between the frame of the robot base and the the frame of the depth camera,the self-made depth camera is used to obtain three-dimensional information of the workpiece scene,and point cloud matching and pose estimation algorithm are performed with the template point cloud of the workpiece to identify the pose of the workpiece,combined with the end fixture of the robotic arm,so the task of grasping workpiece is completed.First,study the reconstruction of 3D point cloud.This paper uses binocular cameras and coded structured light to obtain point cloud.Camera calibration,stereo calibration,and stereo correction for binocular cameras enable the two cameras to form a head-up binocular stereo imaging,which is convenient for obtaining point clouds through disparity map.In this paper,7-bit positive and negative gray code and 8-step phase shift are used to realize the coding and decoding of structured light.A total of 22 coded structured light patterns are projected onto the surface of the workpiece scene.The gray code value and the phase value are used for stereo matching to obtain disparity map,and the triangulation principle is used to obtain three-dimensional point cloud of the workpiece scene.Second,study the calibration techniques of the robotic arm,including DH parameter calibration and hand-eye calibration.The DH parameter calibration algorithm based on the distance error model is used to improve the position accuracy of the end effector of the robotic arm,and to avoid the errors introduced when converting the three-dimensional point from the frame of the measurement equipment to the frame of the robot base.The hand-eye calibration algorithm based on dual quaternion is used to obtain the transformation relationship between the frame of the robot base and the the frame of the depth camera,which has high accuracy and good robustness.Then,study point cloud matching and pose estimation algorithms.This paper uses point cloud matching and pose estimation algorithms based on point-pair-feature,and makes a series of improvements to the characteristics of random bins in the industrial environment,such as scene point cloud normal direction consistency adjustment,filtering of grasping pose strategy adjustment,angle deviation adjustment caused by rotational symmetry to obtain more ideal pose estimation results.Finally,experiment verification and result analysis are carried out.Experiments include grasping experiments in the simulation environment and the real environment.First obtain the grasping pose of the teaching position,and then obtain the transformation relationship from the teaching position to the grasping position through the pose estimation algorithm,and finally obtain the pose of the end effector of the robotic arm under the grasping position,and use this pose to complete the grasping of the workpiece.This paper constructs an experimental platform to obtain 3D point cloud data on the surface of the workpiece scene,analyzes and optimizes related algorithms,completes DH parameter calibration,hand-eye calibration,point cloud matching and pose estimation,obtains the pose information of the workpiece,and then combines the end fixture of the robotic arm to complete the task of grasping workpiece.Finally,the results of experiments verify the graspoing of the random capacitors.The results of simulation experiments and real experiments show that the robot guidance and positioning technology based on point cloud proposed in this paper can complete the task of random bin picking,and has certain reference significance.