Crowd Analysis by Using Optical Flow and Density Based Clustering

In this paper, we present a system to detect and track crowdsin a video sequence captured by a camera. In a first step, wecompute optical flows by means of pyramidal Lucas-Kanadefeature tracking. Afterwards, a density based clustering isused to group similar vectors. In the last step, it is applied acrowd tracker in every frame, allowing us to detect and trackthe crowds. Our system gives the output as a graphic overlay,i.e it adds arrows and colors to the original frame sequence,in order to identify crowds and their movements. For theevaluation, we check when our system detect certains eventson the crowds, such as merging, splitting and collision