Translator Disclaimer
26 January 2009 Beacon recognition and tracking based on omni-vision
Author Affiliations +
Omnidirectional vision (Omni-vision) has the feature that an extremely wide view can be achieved simultaneously. The omni-image brings a highly unavoidable inherent distortion while it provides hemispherical field of views. In this paper, a method called Spherical Perspective Projection is used for correction of such distorted image. Omni-vision target recognition and tracking with fisheye lens for AGVs appears definite significant since its advantage of acquiring all vision information of the three-dimensional space once. A novel Beacon Model and Omni-vision tracker for mobile robots is described. At present, the research of target model has many different problems, such as outdoor illumination, target veiling, target losing. Specially, outdoor illumination and beacon veiling are the key problems which need an effective method to solve. The new beacon model which features particular topology shape can be recognized in the outdoors with part veiled of the object. In this paper an improved omni-vision object tracking method based on mean shift algorithm is proposed. The mean shift algorithm which is a powerful technique for tracking objects in image sequences with complex background has been proved to be successful for the fast computation and effective tracking problems. The recognition and tracking functions have been demonstrated on experimental platform.
© (2009) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Baofeng Zhang, Yuli Liu, and Zuoliang Cao "Beacon recognition and tracking based on omni-vision", Proc. SPIE 7156, 2008 International Conference on Optical Instruments and Technology: Optical Systems and Optoelectronic Instruments, 715610 (26 January 2009);

Back to Top