Point Feature-based Three-dimensional Matching Algorithm

Stereo matching, which is an important research direction in computer vision field, is a bottle-neck problem in the research of stereo vision. It's a great challenge to investigate a robust, precise, stable, and applicable stereo matching algorithm, which is both theoretically and practically valuable.In the thesis, we investigated a number of important problems related to stereo matching problem. The main work can be summarized as follows:1. Corner Detection Technique: We made deep study on the corners detection algorithm based on the gray-level image. A new corner detection algorithm which combines Harris algorithm and Forstner algorithm was given. Contrast with the point's validity, the number of points and the extraction time. Experiment results showed that the improved algorithm had better stability, and the corner points extracted more precise, more uniform distribution.2. Feature Matching Technique : In view of the matching efficiency and performance deficiencies of the SIFT keypoints descriptors, firstly the use of circular rotation invariance of the key points described in sub-structure, thus the final out of the descriptor structure were both unique and robust, then reduced the search space of matching features from 2D to 1D according to epipolar constraint in stereo vision theory. The matching result is finally obtained in the light of the nearest neighbor matching based on Euclidean distance of SIFT descriptors. The idea of combining neighbor orientation information made a great enhancement on robustness against image noise and various image transformations, offering good reliability for feature matching with location error. The experimental results prove that this algorithm wins both high matching accuracy and good robustness against affine transformation, 3D viewpoint transformation and non-linear image intensity transformation. And it can be applied in stereo vision system which is in demand of high robustness for stereo matching algorithm.3. The design of dynamic link library of matching algorithms: According to the two matching algorithms that were given, designed their dynamic-link library which provided the appropriate package of matching algorithm to the three-dimensional reconstruction and the measurement of the camera to research for the future. This work simplifies the programming process and saves time.