Recognition And Localization Of Work-piece Based On Binocular Stereo Vision And Development Of Robot Grabbing System

With the development of science,The applications of machine vision is expanding and deepening,especially binocular stereoscopic vision system will have a greater impact on the flexible manufacturing system.Binocular stereo vision coordinated flexible manufacturing system can make the system free of the robot teaching part,and make the system more flexible,more flexible,more efficient.According to the actual demand,this article based on the theory of binocular stereoscopic vision,two cameras take pictures from different angles.Through those steps that image preprocessing,target feature point extraction,image matching,target recognition and positioning,three-dimensional reconstruction and three-dimensional information transmission to the robot achieved work-piece gripping.Finally,the whole system and the experimental data are deeply studied and analyzed.1.Two cameras are calibrated by using the calibration method of Zhang in two different environments of MATLAB and Opencv,after then we analyze result.Finally,we studied the influence of illumination on calibration.2.Image preprocessing and artifact feature extraction.First of all,this paper through image processing technology: histogram enhancement,median filter to minimize the interference caused by external environment.Then,this paper combines the advantages and disadvantages of Harris corner detection,SIFT and DAISY,and proposes a method of feature point extraction.Finally,the algorithm is validated.3.First,we obtained the centroid of the left and right graph targets and the affine model between the template graph and the search graph,and then the spatial coordinates of the target center are obtained by the three-dimensional reconstruction principle.Finally the coordinates will be transmitted to the robot to achieve the target crawl.In this article,the research and development of Recognition and Localization of Work-piece and Robot Grabbing System based on binocular stereo vision are carried out,through experimental verification,this system performs operates stabled and the accuracy satisfied the actual requirements.The system achieves the goal of improving the flexibility and agility of flexible manufacturing system.