الفهرس | Only 14 pages are availabe for public view |
Abstract Vision is the most powerful sense. It provides us with a remarkable amount of information about our surroundings and enables us to interact intelligently with the environment, all without direct physical contact. There are many techniques for the reconstruction of 3D shape of an object from its 2D gray scale images, shape from shading, shape from stereo and shape from motion methods which have been considered one of the most important area and have been studied intensively in the computer vision area as a field of artificial intelligence for the several decades. The main focus of this thesis is 3D shape reconstruction which, is one of the important subjects in computer vision and a branch of computer science. This thesis presents a scheme for 3D shape reconstruction from stationary or moving object from 2D gray scale images. In this thesis we introduce a reconstruction approach to the local stereo problem using stereo matching technique. We also introduce a new approach for estimating the shape of 3D object from 2D shade image in terms of approximating the height map by a second Fast Trigonometric Polynomial (FTP). The proposed approach satisfies the integerability condition and provides 3D shape estimation with low computational complexity compared to conventional Fourier transform based methods. The experiments on real images show the approaches ability to improve reconstruction accuracy. |