A novel background extraction algorithm for large motion video

Development in the field of very large scale integration (VLSI), over the past decade flooded the market with video cameras, which increases the installation of camera for surveillance purpose. Due to this huge amount of data is augmented per day, which is either impossible to handle manually or to store for too long period. These create the need to find out an algorithms and methods for automatic processing of collected data, to detect object of interest. Automatic processing of data requires some lower level tasks of computer vision to be performed, which include tasks such as tracking targets, labeling foreground parts and detection of foreground objects. This can be done by the process like foreground tracking and background subtraction. To achieve these tasks, it is vital to extract a pure background frame from the video, which does not have any traces of moving object of the scene. This paper proposed the simple probability based background extraction algorithm, which will extract pure color backgrounds of the surveillance videos. Initial background extracted, by proposed algorithm give the accurate and quick results. With the help of this pure extracted background, detection of moving objects was take place correctly and quickly.