Research On Recognition And Orientation Of Object Based On Binocular Vision And Grasping Of Manipulator

At present,most robots are in a fixed environment and repeatedly complete fixed tasks.In order to improve the intelligence and autonomy of robots,machine vision technology is increasingly used in the field of robots.Among them,the binocular stereo vision technology can give the robot a human-like function and can realize target recognition and positioning.Therefore,using the binocular stereo vision technology to improve the robot's intelligence has become a research hotspot.This article is to study the mechanical arm combined with binocular stereo vision technology.Mainly completed the following research content.The positive kinematics and inverse kinematics of the manipulator.Based on the analysis of the spatial pose description and the homogeneous transformation of the coordinate system,using the D-H modeling method,the kinematics model of the UR5 is established;then,the positive kinematics of the robotic arm was studied to obtain the conversion relationship between the coordinate system of the end effector of the manipulator and the coordinate system of the base of the manipulator;based on the positive kinematics of the manipulator,the inverse kinematics of the manipulator is studied,and the joint angles corresponding to the posture and posture of the end effector of the manipulator are analyzed.Finally,this paper uses the Robotics Toolbox in MATLAB to complete the simulation,verifying the validity of the model and the correctness of the kinematics analysis.Parameter calibration of binocular camera.Based on understanding the imaging principle of binocular camera,a reasonable imaging model is established.In this paper,Zhang Zhengyou's calibration method is used to calibrate the binocular camera parameters.We used the toolbox?calib toolbox in MATLAB to complete the parameter calibration,obtained the internal parameters of the camera,and the relative pose of the left and right cameras.Finally,we conducted an error analysis of the calibration results,and the analysis showed that the calibration results were valid.Target recognition algorithm.First,we studied image preprocessing;then,we studied the target recognition algorithm.This article mainly analyzes the two template matching methods of SIFT and SURF.On the basis of determining the target area where the target object in the image is located using the homography matrix,Compared with the ORB algorithm.According to the experimental results,based on the stability of recognition and the efficiency of algorithm execution,an appropriate algorithm is selected and improved.Finally,the effectiveness of the algorithm is improved by real verification..Determine the location of the target.First,we need to perform stereo correction on the collected image pairs;then we can obtain the centroid position of the target object in the template image by calculation.Further,we can use the homography matrix to obtain the centroid position of the target object in the left image,and then we use stereo matching to find the corresponding pixel in the right image,thereby obtaining the corresponding parallax.On this basis,we can reconstruct the centroid of the target object in three dimensions,and the accuracy of the three-dimensional reconstruction is verified through experiments.Build the system and complete the crawling experiment.The system is mainly composed of binocular camera and UR5.Before the grabbing experiment,we carried out the hand-eye calibration of the eye-in-hand system to achieve the conversion from the camera coordinate system to the robot arm end effector coordinate system.After completing the hand-eye calibration,we conducted a crawling experiment.The experimental results show the effectiveness of the method used in this paper.