Application of Virtual Gate for Counting People Participating in Large Public Events

The concept and practical application of the developed algorithm for people counting in crowded scene is presented. The aim of the work is to estimate the number of people passing towards entrances of a large sport hall. The details of implemented the Virtual Gate algorithm are presented. The video signal from the camera installed in the building constituted the input for the algorithm. The most challenging problem was the unpredicted behavior of people while entering the building. A series of experiments during real sport events and concerts was made. The case of improved organization of people passing is described and the influence on the counting results is shown. The results of the studies are shown and achieved outcomes are discussed.