Based On The Three-dimensional Reconstruction Of Three-dimensional Computer Vision

3D reconstruction is the hot point in Computer Vision. This thesis is based on Marr vision theory, doing research on the problems which include the extraction and matching of feature points in the pair of images, the calibration of the camera parameters and the display of 3D model of object with texture in computer, and succeeding in realizing a 3D reconstruction prototype system. The experiment results show that the method used in this article is robust and the effect of reconstruction result is satisfactory .In the computation of fundamental matrix, a new algorithm based on genetic algorithm is proposed and the attempt to get robust solution to fundamental matrix makes sense; On the matching stage, Harris corners are extracted and taken as the feature points . After three steps - correlate matching, relaxation iteration and LMedS, a robust set of match points is gained; The camera calibration uses the Zhang's planar method and the operation is easy and applicable; At last, SFM algorithm is used to make reconstruction from motion, that is Structure from Motion, then the motion parameters of camera is calculated and the 3D coordinates of the scattered points of target is computed . As the result of the 3D reconstruction, the 3D model of the target is display in computer through OpenGL programming using Delaunay triangulation and texture mapping .