A decision support engine for video surveillance systems

Abstract

Design and implementation of an automated or a semi-automated surveillance system is an active research area. Safety concerns of individuals have accelerated the research for identifying alarming human activities in public spaces like streets, shopping malls, airports, and others. But reliance on human operator for real-time actions can be inappropriate and expensive. On the other hand, demonstrating pragmatic scene of a surveillanced area with the help of a synthetic world can be effective that can guard operator's limitation. In this paper, we design a Decision Support Engine (DSE) coupled with a synthetic space to facilitate surveillance activities of operators. For this purpose, the detailed sensory data are processed and alarms are detected, classified and ranked according to the threat severity that follows some well-defined rules of the system. At the end, the identified alarming cases are marked and presented with the help of a synthetic environment in an elegant way so that the operators can take the right action in a specific security circumstance.