3D vision location is one of the most famous researches in computer stereo vision area. However, the traditional algorithm is not only complicated but low-matched. We proposed a new algorithm in this article: First of all, we use Zhang Zhengyou camera calibration methods for vision calibration. Secondly, Epipolar constraint, NCC similarity measure and disparity gradient compatibility are individually used in different stages of sparse feature based stereo matching. The last, least squares space coplanar approximation is used to obtain the 3D coordinates of matching points. It has been proved not only operated easily but provided with high stability and robustness.
Published in:
Intelligent Human-Machine Systems and Cybernetics (IHMSC), 2012 4th International Conference on
(Volume:2
)
Date of Conference: 26-27 Aug. 2012