Detecting background setting for dynamic scene

Abstract

Processing Real-Time image sequence is now possible because of advancement of technological developments in digital signal processing, wide-band communication, and high-performance VLSI. With the developments in video technology, the surveillance system can be built with some low cost gadget such as the web-camera. In this modern life with increasing number of crime rate, people in society need for security and safety; video surveillance has become important reason to oppose threats of crime and terrorism. The most fundamental part of surveillance is foreground detection, which is retrieval of an object of interest. The object of interest can remodel by common background subtraction technique. There is some problem arises by using this technique, where because of variation of light source, the background constantly changes. The intensity of pixel changes throughout the object detection takes place. Intensity of pixel value changes leads to improper foreground detection, the background detected as foreground object. This paper proposes a method to model and update the background of the scene by intersection solving method.