Research And Implementation Of Computer Stereo Vision's Algorithm

Stereo matching is one of the most active research areas in computer vision. Stereo research has matured significantly throughout the last decades and many advances in computational stereo continue to be made, allowing stereo to be applied to real-time systems. Stereo system can be applied in many areas, such as vehicle and robot navigation, video surveillance and manufacture process control. These applications always require high-accuracy dense disparity maps in real-time and online. These local approaches are generally outperformed by offline global optimization based algorithms. However, recent research shows that, through carefully selecting and aggregating the matching costs of neighboring pixels, the disparity maps produced by a local approach can be more accurate than those generated by many global optimization techniques.Most stereo systems work in three steps: camera calibration, stereo correspondence and 3D reconstruction. The foundation of epipolar geometry rectification and 3D reconstruction is camera calibration. The precision of camera calibration determine the performance of stereo system. This paper employs Zhang's approach of camera calibration and advances the precision of result of homogeneity equation. The arrangement of cameras can reduce the complication of stereo correspondence algorithm.Stereo correspondence algorithm is the most important and complicated part of stereo systems. Under the real-time constraint of video surveillance, improve the local approach for effectively distinguishing objects horizontally.A real-time stereo system is established by implementing on windows and visual C++. Analyze the result of camera calibration affected by the number of boards. We provide comparative analyses on the efficiency of our algorithm, dynamic programming algorithm and graph cut algorithm.