期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2012
卷号：9
期号：2
出版社：IJCSI Press
摘要：Detecting stationary objects such as an Abandoned object or a Parked vehicle is very critical and has extensive series of applications. This paper presents a new method for detecting stationary objects in video sequences of traffic taken from a static camera. It uses Running average to model the background image at every pixel position. The background obtained is then used to detect the stationary foreground object in the video. Stationary foreground object can then be classified as illegally parked vehicle or an abandoned object based on the amount of time it has been stationary. The system has been evaluated using public datasets and has been robust to high noise levels. Potential applications include Urban Traffic Surveillance and Incident detection on roads.
