Abstract

One of the trends in computer vision is development of systems using multiple vision sensors. Especially, use of multiple omnidirectional vision sensors (ODVSs) opens up a new application area of computer vision with their wide range of view. As an application of such vision systems, we propose a real-time human tracking system using multiple ODVSs. The system measures people's locations by N-ocular stereo, which is an extension of trinocular stereo, from omni-directional images taken with the ODVSs. N-ocular stereo can solve the correspondence problem among multiple targets without visual features. In addition, several error compensation methods are introduced for accurate localization of targets in N-ocular stereo. We have evaluated the proposed methods in the experimentation and con rmed that the system can robustly track people in real time. 1

Citations