A Distributed Sensor Network for Real-Time Multisensor Surveillance Against Possible Terrorist Actions

摘要

In this paper, a Distributed Sensor Network (DSN) architecture has been designed for fusing multisensor data in a video surveillance system capable of understanding possible terrorist actions. The system uses both physical (i.e., IR and optical cameras) and virtual (i.e. processing units transforming information) sensors. Internal nodes of the DSN fuse data in order to reduce the degree of uncertainty naturally associated with the acquired information, and to produce an interpretation of the environment observed. The performance of the cameras is automatically assessed to regulate the fusion process accordingly. In particular, the system is able to detect, recognize, and track moving people in outdoor environments by analyzing their behavior through trajectory and event analysis. Suspicious behaviors are then signaled to a remote operator supervising the system.