Study On 3D Recongstruction Techniques Based On The Gray Information Of Image

In recent years, with the developing of the communication technology and the computor vision theory, the technology of 3D passive reconstruction has become an important research aspect in computer vision domain. As an important method of passive-acquired 3D data in computer vision, the image based 3D reconstruction techniques can reconstruct the 3D shape of the object from the gray information of a 2D image. Owing to its excellent characteristics, such as easy to operating, low cost, high efficiency and wide area of application, it attracts great attention. In this dissertation, based on the principle of photometry, combined the theories and techniques of Computer Vision, Image processing and CAGD, the principles and realization methods of 3D reconstruction techniques based on gray of image are discussed. The main content as follows:In order to improve the precision of 3D reconstruction, the image pretreatment methods are researched, the principle and realization methods of image pretreatment technique, such as grayscale processing, grayscale transformation, and image improving et al, are given.The question of the 3D data'resolution is limited by the resolution of image was solved by the image interpolation based on the B-spline. Aiming at the character of pixel'scattor, a method of parameter mapping was proposed. The proposed method can efficiently reserve high frequency and detail message of interpolated image.Base on the theory of photometry, the principle of shape from shading (SFS) method is analyzed. Using Phong illumination model, the realization method of minimization approach of SFS is deduced.Based on the above-mentioned theoretical research, under the Windows operation system, using Visual C++ 6.0 and Open GL programming language, a a image gray based 3D reconstruction prototype system is designed anddeveloped. The application effects of the system are verified through experiments, such as simulation of natural scenery, recovering of cultural relics and arts and human face modling. The results of experiments demonstrate that the research method and the algorithms given in this dissertation are practical, feasible and accurate.