Real time person tracking and behavior interpretation in multi camera scenarios applying homography and coupled HMMs

Abstract

Video surveillance systems have been introduced in various fields of our daily life to enhance security and protect individuals and sensitive infrastructure. Up to now they have been usually utilized as a forensic tool for after the fact investigations and are commonly monitored by human operators. A further gain in safety can only be achieved by the implementation of fully automated surveillance systems which will assist human operators. In this work we will present an integrated real time capable system utilizing multiple camera person tracking, which is required to resolve heavy occlusions, to monitor individuals in complex scenes. The resulting trajectories will be further analyzed for so called Low Level Activities , such as walking, running and stationarity, applying HMMs, which will be used for the behavior interpretation task along with motion features gathered throughout the tracking process. An approach based on coupled HMMs will be used to model High Level Activities such as robberies at ATMs and luggage related scenarios. © 2011 Springer-Verlag.