Research On Grasp Positioning Technology Of Industrial Robot Based On 3D Point Cloud

In recent years,flexible industrial robots assisted by vision sensors have assumed more and more tasks in automatic plants.Vision sensor guides the robot to carry out operations such as target grabbing,parts assembly,rejection,etc.by acquiring the information of three-dimensional morphology,spatial posture and surface texture of the target in the field environment.Based on the project "White Sugar Pack Grasping Planning of Palletizing Robot Based on Machine Vision" to improve the quality and efficiency of the flexible robot grabbing system,this paper studies an industrial robot grasp positioning method based on 3D point cloud.Specific research contents are as follows:A calibration method of visual sensor based on point cloud with color and depth information is proposed.A three-dimensional vision sensor consisting of a color camera,an infrared transmitter and an infrared camera is constructed to obtain the surface texture and depth information of the target simultaneously.Based on these two kinds of information characteristics,the 3D point cloud data features with color and depth information are defined,and a joint external parameter calibration method is proposed.A method for calculating the pose of target space based on the color and depth information of point cloud is proposed.Aiming at the characteristics of 3D point cloud data containing color and depth information,firstly,the Region Growing method based on color is used to segment the plane point cloud,then the non-planar point cloud is detected by Random Sample Consensus method,and the cylinder model is extracted.Then the 3D pose of the target is calculated on this basis.An improved Rapidly-Exploring Random Tree method is proposed to realize the motion planning of the robot.Robot hand-eye calibration is used to obtain the transformation relationship between vision sensor and robot.The rapid search random tree(RRT)method based on sampling is used to realize the motion planning of the robot in MoveIt!.The experimental platform is built with UR5 robot as the carrier,and the feasibility and effectiveness of the above method are verified by the experiment of target grasp positioning.The system has good scalability and can be applied to robot grasping tasks in automated factories.