Research On The Detection And Gripping System Of Electronic Components Based On Machine Vision

Along with the rapid development of modern information technology and the Internet,people’s demand for information and communication technology is increasing,which also promotes the application of electronic devices more and more widely.Today,electronic devices have penetrated into the communication,entertainment,medical and industrial production areas of people’s lives,becoming an integral part of modern society,which also poses higher demands on the electronic components produced by the electronics manufacturing industry in terms of quality and efficiency.But with the development and application of semiconductors,materials and other technologies,the speed of electronic components is increasing,making many types of electronic components with higher integration and smaller size,the traditional industrial production methods have been unable to meet the requirements of electronic components production.In this context,a machine vision-based electronic component inspection and gripping system is proposed.By combining machine vision technology with industrial robotic arm gripping and applying it to the industrial production environment in the fields of sorting,welding,assembling,mounting and recycling of electronic components,the automation and intelligence of electronic component production is realised,further improving the production efficiency and production quality of electronic components.The following aspects are studied and analysed in the subject:(1)The general framework of the system is designed for two parts:electronic component detection and robotic arm gripping.In addition,the D-H model parameters of the ABB IRB120 robot arm were constructed and analysed in terms of forward and inverse kinematics by means of the kinematic principle of the robot arm.The camera calibration experiment eliminates the imaging distortion and obtains the internal and external reference matrices,thus completing the conversion from the pixel coordinate system to the base coordinate system of the robotic arm.(2)The visual inspection algorithm using YOLOv5 for robotic arm grasping is proposed based on the production requirements of electronic components in an industrial environment,and its structure and functions of the network are introduced.To further improve the detection performance of the model,three improvements to the model are proposed in terms of Backbone lightweighting,gradient information triage and multi-scale weighted feature fusion.Then an experimental environment is built,a dataset of electronic components is created and the annotation of electronic components is completed.Finally,the convergence analysis of the improved model is carried out and the three improvements are validated by ablation experiments.(3)A joint simulation platform of Move It! and Gazebo was built based on the ROS system for visual inspection and grasping motion of electronic components by a robotic arm.In the process,the system model for Gazebo simulation was created,the ROS visual inspection module with python script was written,the Move It! planner was set up to complete the grasping motion of the robot arm,and the Rviz visualisation platform for monitoring and adjusting the motion parameters of the robot arm was built.Finally,a gripping experiment was conducted and the experimental results were analysed to verify the feasibility of the system simulation.(4)A machine vision-based electronic component detection and grasping experimental platform was built.Based on the target detection network,the electronic components to be grasped were identified and their positioning information was transmitted to the robot arm control cabinet to control the robot arm to grasp the electronic components.The realistic feasibility of the machine vision-based electronic component detection and grasping system was evaluated in terms of recognition accuracy,positioning error and grasping success rate.