Research On Flexible Clamping Robot Workpiece Location Based On Binocular Stereo Vision

With the development of machine vision technology,binocular stereo vision has become one of the trends of machine vision.Compared with monocular vision,the exact three-dimensional information of the target object in a complex environment can be obtained by binocular stereo vision.So the binocular stereo vision was widely used in robot recognition and position.The binocular stereo vision system was established in this thesis to investigate the camera calibration for stereo vision,image feature extraction and matching algorithm,positioning of the target workpiece and 3D reconstruction.The experiment platform of robot workpiece positioning system was also built to complete the accurately positioning of the target workpiece.The main research works of this thesis are as follows:Firstly,by analyzing the advantages and disadvantages of the traditional calibration method and the self-calibration method,a camera imaging model was proposed,then the effects of camera distortion in the imaging process were considered.In order to obtain distortion and camera parameters,a method combining Zhang Zhengyou's calibration method and Heikkila distortion model was proposed.The experimental results showed that,the more comprehensive distortion coefficients were obtained by using the proposed method.Secondly,in order to address the issue of the poor real-time performance of traditional SURF algorithm,an improved SURF algorithm was proposed.The binary descriptor,BRIEF,was used to describe the features,which reduces the complexity of the traditional SURF algorithm.The experimental results showed that,the matching speed has been improved in a certain degree compared with the traditional SURF algorithm,and at the same time,the robustness of the traditional SURF algorithm can be ensured.In the situations of rotation changes,scale changes,illumination changes,image compression and noise,exit in the workpiece images,a high matching accuracy also can be obtained by using the proposed algorithm.Thirdly,for the positioning of the target workpiece,a stereo matching positioning algorithm based on the improved SURF features was proposed.The improved stereo matching algorithm was utilized in stereo matching between the target artifact images,captured by the left and right cameras,to obtain the image coordinates of the effective matching point pairs.The image coordinates of the target workpiece position wereobtained by averaging the datasets of the real matching points.Then the three-dimensional coordinates of the target workpiece position were obtained by combining the binocular camera calibration results and the point-based three-dimensional reconstruction method.The error of positioning was analyzed by experiments,and it was verified that the positioning method has certain application value in robotic workpiece positioning.Finally,the hardware and software systems of the robot positioning system for flexible clamping robot were introduced,and the software operating system was designed based on the VS2010.