Evaluation of disparity estimation techniques and its application in mapping

This paper presents an evaluation of disparity estimation algorithms for mapping in mobile robot navigation context. The disparity estimation techniques compared in this paper are: Block matching and stereo matching based on chain code vectors. The parameters compared in this work are computation time and accuracy. Results indicates that the chain code vector matching technique is suitable for real-time applications as the computation is faster and is advisable for feature tracking as the memory requirement is lesser. The paper also presents the application of these techniques for mapping.