Research And Implementation On 3d Information Reconstruction Based On Monocular Vision

Estimation of the camera motion and structure of a scene from two-dimensional images taken by a camera is one of the core problems in the area of computer vision, which is called Structure from Motion (SfM). Reliable algorithms for motion and structure estimation are invaluable in visual robot navigation, cartography, aerial reconnaissance, target identification and virtual reality in computer graphics.Distinguished from stereo algorithms based on a traditional binocular stereo visual system or multi-view stereo visual system, the algorithm described in this thesis is based on monocular vision, the advantage of which is that it greatly decreases the equipment requirements and can be applied to an uncalibrated amateur handheld camera.This paper addresses the problem of estimating camera's three-dimensional motion and recovering scene structure from motion based on points as long as a projective camera and a rigid reconstruction scene are defined. To recover scene structure from motion, firstly the relation between sequential images must be set up through three steps as feature point extraction, corresponding matching and estimation of fundamental matrices. A method base on density degree of feature points'distributing to estimate image corresponding is presented, which reduces wrong correspondings and improves accuracy of estimated fundamental matrices. In the stage of calibration, we use the method which was presented by Zhang Zhengyou, to calibrate the single camera. Then calibration will be used to determine the internal parameters of the camera in order to recover the camera motion and upgrade the projective structure to metric structure. Finally, a scene reconstruction is set up using matrix decomposition method and triangulation method. It shows experimental results to demonstrate the effectiveness of the proposed methods.