Distributed video data fusion and mining
Edward Yi Chang, Yuan-Fang Wang, Volkan Rodoplu · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2004
This paper presents an event sensing paradigm for intelligent event-analysis in a wireless, ad hoc, multi-camera, video surveillance system. In particilar, we present statistical methods that we have developed to support three aspects of event sensing: 1) energy-efficient, resource-conserving, and robust sensor data fusion and analysis, 2) intelligent event modeling and recognition, and 3) rapid deployment, dynamic configuration, and continuous operation of the camera networks. We outline our preliminary results, and discuss future directions that research might take.