The results of earlier studies on the possibility of spatiallocalization from panoramic images have shown good prospects forview-based methods. The major advantages of these methods are a widefield-of-view, capability of modeling cluttered environments, andflexibility in the learning phase. The redundant information captured insimilar views is efficiently handled by the eigenspace approach.However, the standard approaches are sensitive to noise and occlusion.We present a method of view-based localization in a robust frameworkthat solves these problems to a large degree. Experimental results on alarge set of real panoramic images demonstrate the effectiveness of theapproach and the level of achieved robustness
展开▼
