AUTOMATED VIDEO SURVEILLANCE FOR ACCESS CONTROL

In this paper, the authors present an automated video surveillance system based on dynamic image sequences analysis. They first introduce a motion detector which makes use of the difference between a time varying background model and the edges of the current image of the scene in order to detect intruders in an area under control. The main advantage of the proposed detector is its robustness with respect to illumination changes. Then, they present the integration of the system in a specific tool dedicated to the surveillance of car parks. This tool is used for the PRAXITELE project, a concept under development in France for a novel transportstion system based on a fleet of small public electric cars. The performance of this new video surveillance system is illustrated on image sequences under several environmental conditions.