Moving Object Detection and Tracking in Traffic Surveillance Video Sequences

In the field of traffic video surveillance systems, multiple object detection and tracking has a vital role to play. Various algorithms based on image processing techniques have been used to detect and track objects in video surveillance systems. So it is required to develop an algorithm which is computationally fast and robust to noisy environment. In this paper, variance-based method for multiple object detection and tracking of vehicles in traffic surveillance is proposed and compared with the existing method. This new method is based on five-frame background subtraction, five-frame differencing, and variance calculation for object detection and tracking. To evaluate the proposed method, it is compared with the existing methods such as standard mean shift method. The experimental results show that this method gives better accuracy and takes comparable computational time when compared with mean shift method, which is required in traffic surveillance systems. The comparative shows that the proposed method gives about 99.57% accuracy in detection whereas mean shift gives 88.88% accuracy.