Research On Vision Based Dynamic Target Acquisition Of Robot Arm

Machine vision has always been a research hotspot in the field of vision.As an important part of the field of machine vision,binocular vision system is widely used in various fields of production and life with its rich functions and simple structure.In this paper,the XB7 robot arm is taken as the research object,and the vision based robot arm dynamic target capture system is studied by combining the robot arm kinematics,binocular vision,image processing and other technologies.The main research contents are as follows.First,the overall composition of the system is given,the main hardware facilities in the system are selected and analyzed,and the model and parameters of the equipment are determined.This paper introduces the software platform needed to realize the system functions,and describes the work flow of the system.The pose and three homogeneous coordinate transformations are described in detail.The XB7 manipulator is modeled according to the D-H parameter method,and the forward and inverse kinematics of the manipulator are analyzed.Secondly,the imaging model and distortion model of the camera are analyzed.A design method for the position distribution of the calibration plate with the idea of uniform design is proposed.Using this method and Zhang Zhengyou calibration method,the calibration experiment of the binocular camera is completed in the Matlab image processing toolbox,and the internal and external parameters and distortion parameters of the camera are obtained.The imaging model of the binocular vision system and the ranging principle are introduced in detail.Then,the image of the target object is processed.Firstly,median filtering algorithm is used to preprocess the image;Then the object and the background are segmented by setting a color threshold;Then open the binary image after threshold segmentation to smooth the object boundary.Then,using the boundary tracking algorithm and geometric moments,the boundary contour and centroid of the object in the image are detected to determine its shape and centroid;Then,the object is reconstructed using binocular vision system to determine its three-dimensional coordinates;At last,Kalman filter is used to predict the trajectory of the target object based on the historical motion information of the target object,and the prediction results conform to the actual motion state of the target object.Finally,an experimental platform for dynamic target acquisition of the manipulator based on vision is built,and the hand-eye model of the manipulator and vision system is established,and the hand-eye calibration is completed by establishing the workpiece coordinate system.Then the accuracy of 3D reconstruction is verified,and the verification results show that the error range of 3D reconstruction in this paper is below ± 5mm,which can meet the working requirements of the system.Then the system test experiment is carried out,and the experimental results show that the system can better achieve the acquisition of dynamic targets when the moving maximum height of the target is greater than 3.5m.