Agent Based Object Recognition System for Autonomous Surveillance in Dynamic Environments

Autonomous Surveillance is an important term in order to produce pervasive, ubiquitous, homenet, telemetics and other application purposes. However, many surveillance systems are annoyed with some environmental hazards like illumination and others. This paper presents a novel method for non-intrusive biometric vision system for the surveillance having the prior knowledge about environment. As an environment can change or alters its situation dynamically e.g. strong sunlight, evening, cloudy or rainy, it is very important to be aware or sense about the current environmental situation properly in order to apply predefined actions for each situation. This dynamic sensing environment system can be achieved through the use of multiple real and virtual agents. The proposed dynamic sensing environment system is implemented on an experimental prototype setup for non-intrusive biometric vision system.