Correlative Technology Research On 3D Modeling Based Computer Vision

This dissertation discusses the technique of 3D modeling based on computer vision. The sensor model is set up, and the algorithms of image processing, feature detection and stereo matching are realized.The theory of using binocular sensor to get range information is discussed in the dissertation. The algorithms of image processing, including pixel gray-level detection, intensity transform, noise smoothness and image sharpen are discussed in detail. Next, edge detection using differential operator, edge thinning and connecting are introduced. Model and methods of image segmentation are discussed subsequently. Stereo matching methods, such as correlative matching algorithm and least squares matching algorithm are given and realized.Realization of all algorithms in the dissertation use integrated programming of Matlab and C++.Finally, the review and further research of this dissertation are pointed out.