Research On Positioning And Grabbing Of ABB Robot Based On Vision Guidance

At present,with the continuous improvement of the level of industrial automation,the demand for industrial robots is increasing.At the same time,the rapid development of computer technology and image processing technology,combining visual guidance technology with industrial robot technology,can greatly improve the flexible operation and production efficiency of industrial robots.In order to solve the problem of industrial robots sorting different workpieces,this paper will combine computer image processing technology to design a set of target guidance and positioning visual guidance system to carry out experimental research on the above problems,and improve the intelligence and flexibility of industrial robots.The main research contents of this paper are as follows:(1)The design of the vision guidance system.In this paper,the target object needs to be identified and positioned by the industrial camera.Therefore,according to the experimental environment and the object to be tested,the hardware of the vision guidance system is selected,the flow and scheme of the whole vision guidance system are designed,and finally the experimental platform is built.(2)Perform camera calibration and hand-eye calibration of the vision system.Determine the connection mode between the camera and the robot as the model outside the hand.Use the HALCON visual library to calibrate the camera of the monocular camera to obtain the internal and external parameters of the camera.The calibration model of the robot and the camera is performed by using three different calibration plates.The hand-eye calibration determines the conversion relationship between the pixel coordinate system and the robot coordinate system.(3)Study the problem of identification and location of multiple target workpieces.In this paper,image preprocessing is performed on the acquired image data,the target object is extracted and the pixel coordinates of the center point are calculated,the transformation matrix of the pixel coordinate system and the robot coordinate system is verified,and the transformation matrix with the smallest positioning error is selected.The MLP neural network model is used to The target workpiece is classified and combined with the conversion matrix of the hand eye calibration to determine the position of the target object in the robot coordinate system.(4)Analysis of the forward and inverse kinematics of ABB robots.The D-H parameters of the robot are established.The mathematical model of the forward and inverse kinematics of the robot in the initial state is verified by Matlab,so that the gripping posture of the robot can be obtained,which provides guidance for determining the trajectory of the robot.(5)Finally,the sorting experiment verification of multiple target workpieces is carried out,and the experimental results are analyzed: this vision guiding system has a good effect on the recognition and positioning of multiple target workpieces.When the camera resolution is 1280×640,The position error is about 0.742 mm,and the rotation angle error is 0.59°.The robot can perform millimeter-level sorting operations to meet the technical requirements of the system.