Research On Online Hole Detection Method Of Robot Drilling And Riveting Based On Binocular Vision

With the continuous development of industrial automation,machine vision technology has been widely used,and its advantages in hole shape detection and workpiece recognition have become more prominent.In the assembly of various holes,the quality of the hole shape directly affects the assembly accuracy.For example,the hole position and hole shape accuracy on the aircraft riveting parts directly affect the aircraft's assembly quality and flight safety.Therefore,to solve the problems of insufficient intelligence and low detection accuracy of hole detection methods in the current drilling and riveting process,this paper proposes a hole shape and position online detection method based on binocular stereo vision technology.Compared with traditional detection methods,binocular vision detection can obtain the three-dimensional information of the target object more flexibly,which has important research significance.This method uses the AUBO-i5 lightweight 6-DOF robot as the motion body,and the end of the actuator is equipped with a binocular vision system to complete the online detection of hole shape parameters.The main research contents and results are as follows: ?First,This article elaborates on the current status of automatic drilling and riveting technology,drilling,and riveting detection technology,and binocular stereo vision technology at home and abroad,according to the overall design indicators and design principles of the system,hardware selection,and installation,etc,completed the vision system the hardware structure design.Based on the pinhole imaging model,the conversion relationship between the camera imaging model and the imaging coordinate system is established.Through the comparison of the experimental data of the two calibration methods of Matlab and Halcon,the Matlab calibration is selected based on the calibration accuracy to obtain the internal and external parameters of the camera;the binocular is studied.In the calibration and correction of the camera,the rigid transformation matrix of the system is obtained by hand-eye calibration,and the transformation from the camera coordinate system to the end effector coordinate system is realized.Secondly,in terms of three-dimensional reconstruction of riveted holes.First,research on the image preprocessing algorithm is carried out according to the actual working conditions,combined with the image filtering and histogram equalization algorithm to obtain a better processing effect.On this basis,the stereo matching algorithm is analyzed and compared,and SGM stereo matching is selected.The algorithm obtains the disparity map of the scene and realizes the 3D reconstruction of the target hole position based on the disparity map.Thirdly,based on the moving least squares method,this paper has researched the data fitting algorithm,and given the deficiencies of the algorithm itself,completed the improvement of the moving least squares algorithm.It is verified through simulation that the improved algorithm can effectively improve the fitting accuracy.The system realizes the correction of the normal pose based on the improved moving least square method.Finally,an experimental system was built based on the AUBO-i5 robot and the Haikang MV-CA060-11 GM industrial camera.In the experiment,a standard hole designed to simulate the wing is used as the inspection object,and the binocular vision inspection method in this paper is used to obtain the hole diameter and hole position information.Experimental results show.