Abstract：This paper presents a novel localization method based on motion vision. The proposed method selects two feature points from the target to which a mobile robot moves and determines the poses, relative to the target, of the robot before and after motion according to the image coordinates of the two points. A search algorithm is further presented to find more correctly the image coordinates of the feature points. Experimental results indicate that this method has high localization precision and good robustness.
[1] Gehrig S K, Stein F J. Dead reckoning and cartography using stereo vision for an autonomous car[A]. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems[C]. South Korea:1999. 1507-1512. [2] Yagi Y, Nishizawa Y, Yachida M. Map-based navigation for a mobile robot with an omnidirectional image sensor COPIS [J]. IEEE Transactions on Robotics and Automation, 1995, 11(5):634-648. [3] Oberkampf D, Dementhon D F, Davis L S. Iterative pose estimation using coplanar feature points [J]. Computer Vision and Image Understanding, 1996, 63(3):495-511. [4] Hansye S D, Anil K J. Mobile robot localization in indoor environment[J]. Pattern Recognition, 1997, 30(1):99-111. [5] Mouaddib E M, Marhic B. Geometrical matching for mobile robot localization [J]. IEEE Transactions on Robotics and Automation,2000, 16(5):542-552. [6] Huber J, Graefe V. Motion stereo for mobile robots [J]. IEEE Transactions on Industrial Electronics, 1994, 41(4):378-383. [7] Hung Y S, Ho H T. A Kalman filter approach to direct depth estimation incorporating surface structure [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1999, 21(6):570-575.
