The Key Technologies’ Research On3D Reconstruction Based On Binocular Vision

The three-dimensional reconstruction is one of key technology in computer vision. Image reconstruction is to obtain useful signal from images and get three-dimensional information of the measured object. With the rapid development of the computer technology, three-dimensional reconstruction of image has become a a research focus in the field of computer vision and is widely used in aerial mapping, visual navigation, motion analysis, medical diagnosis, industrial inspection and other fields. Therefore, the research in this paper is of theoretical significance and practical value.On the basis of the analysis of a number of related technology and literature, camera calibration in three-dimensional visual field, corner detection, image matching,3D visualization and3D reconstruction are researched in this paper.The main contents are as follows:(1)In this section, the principle of the camera calibration and some classic algorithms are researched. The purpose of camera calibration is to establish an effective imaging model, and to determine the internal and external parameters of the camera, which will obtain the correct relation between the points in the space and the pixels in the image.(2)The SUSAN corner detection algorithm and the Harris corner detection algorithm are researched. According the merits and demerits of these two algorithms, combined with multi-scale and SUSAN algorithm, an improved Harris algorithm is put forward. Through many tests, the validity of this algorithm is confirmed with better effect of corner detection and noise immunity. (3)The classic Harris algorithm can obtain the accuracy of the pixel level, this paper proposed an improved algorithm of sub-pixel comer detection accuracy, which achieved sub-pixel comer detection based on the Harris algorithm, thus improving the accuracy of feature matching.(4)SIFT algorithm is researched, combined the Harris comer detection and the merit of high accuracy of SIFT, Harris-SIFT algorithm for comer detection is proposed.(5)In the combination of stereo camera model and comers matching, the three-dimensional coordinates of points are calculated, and then reconstructed by MATLAB.