【摘要】 根据对月球车立体视觉系统及其工作环境特点的分析,提出了一种特征辅助的区域匹配算法。该算法将匹配分为特征匹配和区域匹配两个阶段进行。特征匹配阶段利用几何约束和多特征匹配得到高可靠的边缘匹配结果。边缘匹配结果经滤波后用于辅助区域匹配,有效地减少了无特征辅助的区域匹配在无纹理和少纹理区域容易产生的误匹配。算法结合特征匹配和区域匹配各自的优点,可以以较高的可靠性得到密集的视差图。实验结果表明,该算法用于室外自然地形图对的匹配时,可以生成较为准确和密集的地形图,满足月球车立体视觉导航的需要。更多还原

【Abstract】 A feature-assisted area-matching algorithm for lunar rover is brought forward based on the analysis of the stereo vision system and working surroundings of lunar rover.The matching is divided into feature-matching phase and area-matching phase.In feature-matching phase geometric constraints and multi-criterion matching strategy are used to build highly reliable edge matching results.Edge matching results are then filtered and utilized to assist area-matching and to reduce the amount of false matches occurring in texture-lacking areas in ordinary area-matching.The algorithm integrates the merits of feature-matching and area-matching and is capable of building dense disparity map with rather high accuracy.Experimental results indicate that the algorithm performs well when used in the matching of stereo image pair of outdoor terrain and satisfies the need of vision-based navigation for lunar rover.更多还原