An effective Traffic Monitoring video segmentation method

With the abrupt increase of motor vehicles, the advanced Intelligent Traffic Monitoring System (ITMS) has become a social hot spot in today´s world. Automatic video segmentation plays an important role in the ITMS. Up to now, researchers have proposed several algorithms on automatic video segmentation. However, because of the large computation, many algorithms cannot be used for the real-time applications. In addition, some algorithms have many parameters to be set in advance. Shao-Yi Chien has proposed a fast video segmentation algorithm, which could meet real-time requirements. In this paper, we proposed an improved algorithm to get more accurate segmentation results from the traffic monitoring videos. With low computation load, the speed of this algorithm, which is running on a personal computer, can meet the real-time requirements. In order to optimize the results of segmentation, we have proposed a post processing algorithm to adjust the coarse segmentation. Experimental results show that our algorithms can give accurate segmentation results without complex computing.