The Research Of Human Face 3D Reconstruction Method Based On Binocular Stereo Vision

As a new field based on various interdisciplinary, the theory research and practical application of the computer vision has made great progress. Especially in recent years, the development of image processing technology and high performance computer coming into being provide favorable conditions for computer vision research and application. So, the further research to computer vision has much theory meaning and practicality value. The study of generating realistic 3D human face models has been a long-time challenging topic in computer graphics. With the development of character animation for films and advertising, computer games, virtual reality, teleconferencing, user-interface, surgical facial planning, and criminal discrimination, etc., this research area is now receiving more attention than ever before. This thesis makes a deeply research on the field.This paper is based on Marr vision theory, expatiate domestic and overseas development and achievement on binocular stereo vision. Based on analysis and conclusion of diversified methods, we studied every part of computer vision systematically.For edge detection, the paper proposed a preferable edge detecting method aiming at binocular stereo vision by comparing these arithmetic. First, do median filter to image in order to wipe off noise, then make use of Gauss-Laplace operator to pick up edge. The result showed the kind of edge detecting method has good performance. For image segmentation, do region partition afresh to original 2D gray histogram, lay the emphases of research on near the area near the diagonal zone, thus this method can not only reduce the operation, but also get rid of the influence of noise effectively. For stereo matching, a feature-based and phase-based method is presented. First of all, we match the edge characteristic points based on feature. Secondly, with the results of feature-based matching as a whole constraint of matching path and phase information, we can solve problems according to whole optimum constraint. This method makes good use of the respective advantages of feature-based arithmetic and phase-based arithmetic. Experiment results showed the high quality and distinct texture of human face modeling in this method.Finally the theoretical analyses and the measurement results showed that the method presented in this paper can partly improve the operation speed and measurement accuracy with simplifying measuring equipment and reducing measuring costs.