Research On Autonomous Sorting System Of Logistics Parcel Based On Vision

In the Internet era,artificial intelligence is rising,and with the rapid development of the logistics industry and China’s aging is becoming more and more serious,the task of manual sorting operations is becoming more and more heavy,resulting in an increase in labor costs,traditional industrial robots can only use a single teaching function for repetitive work,it is difficult to get a wide range of applications,low generalization,there is a great demand for intelligence,the system is based on visual combined with industrial robots,the logistics parcel robot autonomous sorting system research,The main work and research contents are as follows:(1)The robotic sorting system was built using hardware such as ABB IRB4600 industrial robot,Intel Realsense D455 camera,and computer with Ubuntu 18.04.The main operation process of the scheme is: first,the image is collected by the camera,the YOLOv5 s is used to identify and locate the express box on the handling platform,and the target position is converted into the coordinates of the robot at the same time,and then the target object is grasping by the controller control robot through algorithm planning.(2)For the calibration of the internal parameters of the camera,under the ROS system,the internal parameter calibration of the camera is completed by Zhang Zhengyou calibration method,and the conversion relationship between the camera and the robot base is calculated by the Tsai two-step method in the way of eyes outside the hand.(3)For target recognition and 3D coordinate acquisition,YOLOv5 s algorithm is adopted,the dataset is first collected and annotated and the dataset is converted to VOC format,the network model training is carried out after the production of the dataset is completed,and the best weight file is obtained,and after the target is recognized,the distance between the object and the camera is obtained by calculating the bounding box center coordinates of the identified target object and the depth camera,and the position information of the target object under the camera is obtained in combination.After the transformation relationship,the target position information under the robot base is finally obtained.(4)Real-time communication and control between the host computer and the physical robot,in order to realize the control of ABB robot between PC through ROS,the simulation platform is built,first obtain the URDF file of the relevant model robot from the ABB official website,and then configure Move It! in the ROS platform,and the selection of the path planning algorithm is discussed,and the RRT-Connect algorithm is finally selected for the path planning of the robot.At the same time,the robot simulation system is built under the Windows system using Robot Studio software,and the configuration of related ROS communication tasks is completed in the robot controller to complete the real-time control of robot simulation.(5)The construction of the autonomous sorting system platform,the debugging of the visual unit,the testing and analysis of visual positioning,the debugging of the robot unit were carried out,the identification of the express box by the autonomous sorting system was completed,and finally the overall autonomous grasping verification was carried out through the robot autonomous sorting experiment,a total of 48 autonomous grabbing experiments were carried out,and the capture success rate was 93.75%,which verified the feasibility and stability of the system and basically met the expected requirements.