Font Size: a A A

Research On The Industrial Robot Sorting Technology Based On Binocular Vision

Posted on:2022-09-02Degree:MasterType:Thesis
Country:ChinaCandidate:C TianFull Text:PDF
GTID:2518306575463914Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
In the development process of robot and machine vision technology,vision technology has further improved the intelligence and autonomy of robots.At present,most of the sorting systems based on industrial robots are oriented to structured operating environments,resulting in a single sorting objects and low efficiency.In order to enable industrial robots to complete the sorting tasks of multi-category workpieces in an unstructured environment,this thesis studied the industrial robot sorting system based on binocular vision and its key technologies.The main research contents are as follows:Firstly,the overall design of the sorting system was carried out,including the selection of the hardware platform and the selection of the software platform.At the same time,the kinematics analysis of the AUBO-i5 ind ustrial robot was carried out using the D-H parameter method,and the kinematics simulation verification experiment was completed through MATLAB.Then the imaging model of the camera,the principle of dual-target calibration and the principle of hand-eye calibration were described,and the dual-target calibration experiment and hand-eye calibration experiment were carried out using MATLAB and HALCON.In order to facilitate the subsequent extraction of image edge features,this thesis also studied the image preprocessing.Secondly,for the problem that the accuracy of the parallax is easily interfered by factors such as illumination,weak texture,and occlusion in the current process of binocular stereo matching to solve the parallax,this thesis improved the stereo matching algorithm based on the traditional Census transform,and integrated the gradient information of the image into the matching cost function.Therefore,a stereo matching algorithm based on improved Census transform and gradient fusion was proposed to obtain a high-precision disparity map.Then,for requirements that multi-target workpiece recognition and positioning of the sorting system,this thesis used template matching technology to identify and classify multi-target workpieces and determine the grasping position of the target workpiece in the image.Then a method for detecting the pose of the workpiece based on the auxiliary reference line was proposed to determine the posture of the workpiece.And according to the disparity map obtained by the stereo matching,three-dimensional reconstruction was carried out on the grasping position of the workpiece.Finally,in the mixed development environment constructed by Visual Studio 2017,Qt and HALCON,the upper computer interface was written by the C++programming language,and sorting experiments were carried out.The experimental results showed that the angular positioning accuracy of the sorting system in this thesis is up to 7°,the position positioning accuracy is up to 1.6mm,the sorting success rate is higher than 90%,and the system has good accuracy and stability.
Keywords/Search Tags:industrial robot, binocular stereo matching, workpiece recognition, template matching, three-dimensional reconstruction
PDF Full Text Request
Related items