Research On Key Technologies Of Visual Servo For Automatic Loading And Unloading System Of Six-axis Robot

Relying on the existing FMS(Flexible Manufacturing System)flexible manufacturing system and intelligent management and control software,the intelligent manufacturing demonstration line project aims to build a complete intelligent manufacturing system,and plays an exemplary role in efficient automatic production logistics,efficient and safe industrial automation production line,flexible fully automated production system and so on.With the increase of manpower cost and the development of intelligent manufacturing,it is difficult to solve the problems in intelligent manufacturing factories efficiently and controllably by using traditional human or semi-automatic methods.In recent years,the development of machine vision provides more possibilities for the realization of intelligent manufacturing.In the loading and unloading process of the intelligent manufacturing demonstration line project,the operation automation is required.After investigating the working situation of the site,this paper designs a 2D visual grasping scheme to complete the capture of typical workpieces on the AGV(Automated Guided Vehicle)car,designs the hardware composition and function modules of the system,and formulates the technical roadmap.By using a variety of traditional image processing algorithms,the posture recognition of typical workpiece is completed,and the three-dimensional posture of cylindrical bar is obtained.For the case of low contrast and low definition of the original image,the principles of three commonly used image enhancement algorithms are described,and experiments are carried out on the original image under different working conditions.the image enhancement algorithm suitable for this project is determined according to the image visual effect and objective evaluation index.In order to solve the problem of image feature information extraction in the process of visual capture,the mapping relationship between the image obtained by the camera when the end of the robot is in the desired position and the image obtained by the camera when the end of the robot is in the current position is established.the extraction,description and matching of image features are realized.In order to solve the problems of low model estimation accuracy and real-time performance of the traditional RANSAC(Random Sample Consensus)algorithm in the matching process,a quadratic matching strategy and a RANSAC optimization algorithm based on a priori probability sampling are proposed,and experiments are carried out on the enhanced image to verify the effectiveness of the algorithm.In order to realize the visual grasping of typical workpieces,the preliminary modeling work such as kinematics modeling of industrial six-axis robot,calibration of monocular industrial camera and hand-eye calibration of Eye-in-hand configuration are completed.Aiming at the situation that the rectangular workpiece on the AGV trolley may be stacked,using IBVS(Image-Based Visual Servoing)visual servo algorithm,the PID and the controller based on Jacobian matrix are designed,the on-line estimation of the controller is completed,and the control system model is built by Simulink,and the simulation experiments are carried out on the grasping of cylindrical bar and the visual grasping of stacked rectangular workpiece.The results show the effectiveness of using IBVS to grab stacked rectangular workpieces and the effectiveness of the proposed 2D visual grasping scheme.