Research On 3D Reconstruction Of Virtual City Based On The Image Sequence

In recent years, scene reconstruction based on the image sequence is a new method of virtual reality. It is based on the theories and methods of computer vision and image processing, and attempts to use real images to replace the traditional geometric modeling. Also it uses the change of image space instead of the traditional rendering process. Compared to traditional computer graphics, the image-based rendering technology has a lot of unique advantages in constructing realistic scenes and rendering speed. So, it has become a hotspot with the development of Computer Graphics, and has good prospects especially in the construct of virtual environment.3D reconstruction of virtual city is of great significance to the reproduction of the landscape in history,the model expression in Urban planning,the management of urban public facilities,simulation of the building in architectural design and the military and national defense etc. This paper mainly contains the camera calibration,feature extraction and matching,texture extraction,calculation of 3D information and the drawing and roaming of the 3D virtual city.Camera calibration is the basis for 3D reconstruction based on binocular vision. Camera calibration method is discussed in this paper. A new calibration plate is designed for the camera's intrinsic parameters calibration and the experiment is finished. External parameter calibration is finished based on the matching points.Constraint relations between the image sequences are described by the match points. Sequence image matching based on point feature is presented in this paper. Firstly, the Harris corner detector is used to detect corners, a method to describe the direction character of the corners is given and the keypoint descriptor of SIFT(Scale Invariant Feature Transform) is introduced in this method. Secondly, Euclidean distance is performed to get exact matching in the above corner set. Finally, an easy and effective method to avoid wrong matches is discussed. This method is effective on brightness invariant, movement invariant and rotation invariant in the matching of Digital City image sequences. This method is proved to be effective by realistic experiments, and it can be used in practice.The virtual city is of more realistic by using the real texture. Texture extraction method based on the corner orientation is presented in this paper. The texture extraction is orientated by the Harris corner. This method is proved to be accurately and efficiently in the regulation texture extraction.The method of data description means the data structure for the virtual system based on the 3D data and the texture data. A simple building object description method based on the CSG(Constructive Solid Geometry) theory is designed in this paper. And, this method is used in the finally drawing and roaming of the 3D virtual city.