Research On Machine Vision Technology In 3C Flexible Wire Operation

With the continuous improvement of our national living standards,the consumption of 3C electronic products has gradually increased,which has promoted the prosperity and development of the 3C industry.According to public data,our country occupies 70% of the global 3C industry's production capacity.Human cost is one of the main costs of 3C product production,so it is necessary to improve the automated production of 3C products.A large number of flexible wires are used in 3C products.However,flexible wires have the characteristics of high flexibility,difficult to control deformation,and easy to deform,so it is difficult to realize automatic operation.For this reason,this article has carried out relevant research work on the automatic operation of flexible wire in 3C products.In order to solve the difficulties of automatic operation of flexible wires,a vision system based on three-dimensional data is designed in this paper.The system can realize the classification of flexible wires,the establishment of deformation models,and the control of deformation trends,so as to realize simple operations on flexible wires.This paper carries out research on flexible wire classification model,flexible wire deformation law,and visual servo control technology,and builds a flexible wire visual servo control robot system.This paper considers the characteristics of flexible wires and the characteristics of the production environment,analyzes and designs a vision system based on 3D data and a robot visual servo control system,and builds a robot system composed of a six-degreeof-freedom robot,a three-dimensional linear laser camera,a three-dimensional depth camera,and flexible grippers.To realize the functions of classifying common flexible wires in 3C products,establishing deformation models,and controlling deformation trends.Aiming at the characteristics of different types of flexible wires,deformation trends and types of flexible wires,etc.,a flexible wire classification model based on Point Net++is established,point cloud data is preprocessed using PCL library,and point cloud data of flexible wires are collected to train the classification model and verify The classification accuracy of the model.This paper studies the deformation laws of various flexible wires,and establishes a flexible wire shape perception algorithm based on the Correlated Point Drift(CPD)algorithm,which can smoothly realize the real-time perception of the shape of the flexible wire in the presence of outliers,noise,and partial occlusion.On this basis,a robot control system with ROS as the middleware is established to realize data interaction and data synchronization,establish the mapping relationship between the deformation of the flexible wire and the end effector of the robot,and establish the Jacobian matrix of the initial deformation of the flexible wire through the method of finite differentiation.On the basis of the above parts,this paper studies the visual servo control technology of flexible wire rod based on the weighted least squares minimization method with sliding window to optimize the Jacobian matrix online.The corresponding control framework is proposed,the corresponding visual controller is realized,and then the control of the deformation trend of the flexible wire is realized.And through experiments to verify the visual servo control framework.Finally,this paper integrates the above systems and conducts flexible wire operation experiments.Through the simulation of the 3C production line through the simulation environment,the flexible wire operation is completed to verify the effectiveness of the above systems.