Challenges of distributed intelligent surveillance system with heterogenous information

CCTV and sensor based surveillance systems are part of our daily lives now in this modern society due to the advances in telecommunications technology and the demand for better security. These systems are traditionally used in forensic mode – finding evidence in video images when certain events detected or happened. In recent years, research and development on events detection in real time CCTV surveillance has attracted significant attention, in order to identify and prevent potential threats. In this paper, we discuss some challenging issues faced by the artificial intelligence research for such real-time distributed intelligent surveillance systems, where the detection and composition of threats and abnormal behaviors involve multiple sources (e.g., cameras) with heterogeneous information. These challenges include but are not limited to resolving conflicting conclusions provided by different sources; managing uncertainty associated with the conclusions from the sources and the reliability of the sources themselves; managing the heterogeneity of information provided; exploring the scalability and ontological issues presented in large surveillance networks, as well as evaluation criteria etc.