Optical and streakline flow based crowd estimation for surveillance system
G M Basavaraj, Ashok Kusagur · 2016
Recent advancement in computer vision system and growing population in the world has increased the consideration of the computer vision based system for the surveillance application in the real world scenario i.e. shopping malls, sport ground etc. The approach of surveillance requires the tracking of the persons and analysis of their motion. Existing approaches fail to address the problem of tracking and analysis due to the more density of the crowd, more occlusion in the video datasets or scenes. As the density or the occlusions increases, the quality of tracking and analysis also decreases which is a challenging task for the researchers. In this work we propose a new approach for computer vision based surveillance system which considers the video data as input by converting the frames to analyze the motion. We compute the flow and path by computing the streakline of the given frame sequence. This process results in the representation of the flow of the object to recognize the changes in the motion in spatial and temporal domains. This process is gives the better motion segmented results and outperforms the state-of-art techniques for the surveillance system.