利用平面约束的月球车立体匹配方法

为了满足月球车视觉系统对检测障碍物检测的时效性和可靠性的需求，提出了一种基于平面约束和自适应惩罚参数的半全局立体匹配算法。首先，对极线校正后的两幅图像进行SIFT特征点提取与匹配，同时提取边缘特征；然后，利用匹配的SIFT特征点拟合空间平面，并根据平面估计左右图像所有像素点的视差搜索范围；最后，基于传统的半全局匹配算法，采用自适应惩罚参数的策略对左右图像进行立体匹配。实验结果表明：所提出的算法有效地降低了计算复杂度，其计算复杂度只有传统方法的19.9%，对于视差不连续区域以及遮挡区域都能够获得正确的匹配结果。较传统半全局匹配方法无论在速度还是匹配精度上都得到明显提高，为对立体匹配的实际应用奠定了基础。

Stereo matching method based on the plane constraint for lunar rover

To meet the requirements of timeliness and reliability of the lunar rover system for obstacle detection, a semi-global stereo matching algorithm based on plane constraint and adaptive penalty coefficient is proposed. Firstly, the SIFT feature extraction and matching algorithm is used for the two images of epipolar rectification, and meanwhile edges are extracted. Secondly, the matched SIFT feature points is used to fit the plane, and the disparity search ranges of all pixels are estimated respectively in left and right images. Finally, based on the traditional semi-global matching algorithm, the adaptive penalty parameter strategy is applied to the stereo matching of the left and right images. The experimental results show that the algorithm can effectively reduce the computational complexity which is only 19.9% that of the traditional methods, and this method can obtain the correct matching results for the disparity discontinuous region and the occlusion region. Compared with the traditional semi global matching method, the processing speed and the matching accuracy are obviously improved. The method can lay the foundation for the further application.