Smart video surveillance system

Abstract

Smart video surveillance is one of the potential applications of research area of image processing and analysis. Images from video or fixed camera are used as input to a computer vision system. Nowadays, video surveillance has been a key component in ensuring security at airport, banks, offices, governments’ agencies, business and even public areas. In our system, the camera is stationary and the background changes slowly compared to the foreground, so back-grounding is used as a technique for motion segmentation and an easy way to segment the foreground. However, a background model created using simple techniques like averaging will not be able to adapt with changes in illumination, appearance of stable objects, multimodal backgrounds. As the result, background subtraction will not produce accurate motion segmentation in many situations. Therefore, adaptive background modeling can be used to dynamically remodel the background to ensure good results in the presence of background changes. Once backgrounding works well as we expected, tracking and monitoring human processing will be implemented fast and more accurately. The result of the system is very satisfactory and well-applied for practice applications. Because the system is simple in hardware and easy to synchronize with software, it is potential to use widely in ensuring security for preventing from terrorists, criminal problems, and for effective management of public situations.