PNRM MoD Project ADVISOR I (2005-2007) – An automatic visual surveillance system for prevention of terrorism activities and for development of counter measures in case of possible threats
Recent terroristic attacks happened in the world have highlighted the need of the development of new systems able to analyse real situation then prevent dangerous actions that could be acted by terrorists. A continuous supervision to check regular activities is characteristic of public environments like subways and train stations, airports, roads, highways, parking lots and waiting rooms. In the last years, the market availability of data acquisition and processing systems with their low price/performance ratio and more reliable sensors have allowed to increase the number of surveillance applications. Recent achievements in the fields of artificial vision and image processing have driven the development of surveillance systems that are based, in first instance, on information received by different cameras (optical , infrared, acustic, etc..). Automatic image processing and understanding is adopted in order to detect particular alarm condition (suspicious events, suspicious human behaviour, etc…).
Video-surveillance targets vary with the applications needs, could be security as monitoring as control. The actual project propose a system that should operate in complex environments like airports, banks, offices, barracks, etc… and more general in all those contexts in which a control action is necessary to monitor people (or vehicles) accessing and operating in public or risk environments. It should detect suspicious events, recognize people belonging to a set of critic subjects (terrorist or wanted). What this type of systems expects is not only a simple detection and recognition of the objects in the scene but also recognition and understanding of their behaviours.
The proposed system should satisfy different functional requirements: robustness, efficiency and reliability. Robustness is the system capacity to tolerate the random variation of environment conditions that could degrade the quality of the signals and data acquired. Efficiency is an important parameter that allows to evaluate the system time reaction when an alarm occurs. Reliability is given by the correspondence of what the system detect with the real conditions. An outdoor video-surveillance system must be able to act in real-time, to operate in complex environments where light conditions can vary suddenly or mobile objects could be occluded etc.
- Category: National project