Researches On 3D Reconstruction From Uncalibrated Image Sequences

The Structure from Motion (SFM) problem is the central problem in computer vision community. The uncalibrated structure from motion problem (USFM) is the hot topic in the recent 10 years. Many theories, including multiple view geometry, projective reconstruction and camera self-calibration etc. have been proposed and well developed. In the applications, as the technologies of computer and Internet developing, many applications require 3D graphic data of high quality. One challenging problem is how to obtain complicated and realistic 3D graphic models as automatically as possible. This thesis is focused on this problem and improves some key algorithms in the USFM problem. The effectiveness of our algorithms is evaluated in the experiments with simulated data and real image sequences.Firstly, on the research of interest point detection, since when use SUSAN algorithm to detect real image, the number of features is too much, so we propose a improved SUSAN algorithm to decrease the number of corners and to control the density of features in the image. Experiment proves our algorithm can reserve good features and match them profitably;Secondly, in the computation of fundamental matrix, it always turns to optimization problem. An algorithm based on genetic algorithm is researched and the attempt to get robust solution to fundamental matrix makes sense;Finally, after determine the camera's intrinsic calibration parameters. a Euclidean scene reconstruction is set up using matrix decomposition method and triangulation method. It shows experimental results to demonstrate the effectiveness of the proposed methods.