The Research For The Binocular Stereo Matching Based On The Computer Vision

The stereo vision technique is one of the basically classical issues, it can most possibly imitate the human's eyes. So as one of the hotspot, it will be meaningful to make deeper research. Some key factors (such as stereo vision) are profoundly studied. The stereo vision technique mentions that it observes objects from the two different viewpoints, and then gets the spatial information and the contour information under the help of some imaging geometry principles which can get a result by calculation, and then get the result of the subject's 3D information.The binocular stereo matching technology can be divided into the image obtaining, camera calibration, feature extraction, stereo matching and 3D reconstruction. And in the process of the calibration, a special way called Zhang's calibration method which is based on the plane calibration temlate is used, and it ranges from the traditional calibration ways to the self-calibration ways, and then the image coordinate of the subjects' features is obtained by ways of hand working method and the Harris corner detecting, and then the camera parameters can be obtained by the calculation and the experiment. When the stereo matching is working, it first needs to find the respective position of the same subject on the two images, and then find the matching features, decide the matching principle and the algorithm implementation, which are the most important steps on the stereo matching methods.Among the stereo matching methods based on the features, the Sift feature extracting and the matching algorithm are applied, and then the matching points' coordinates are obtained. At last, the 3D coordinate can be obtained according to the basic principle of the stereo vision and the coordinate calculation methods, and then compare the feature spatial information result with the result from the laser sensor, the error will be obtained, then the error analysis is taken, and the contour depth graph is obtained by the spatial point coordinate.The program based of the Sift feature extracting algorithm is compiled under the circumstance of the Matlab and C++.