A Rapid Method for Passing People Counting in Monocular Video Sequences

People flow information is very useful in the real world. This paper presents a novel implementation method of people counting which detects and tracks passing people through an entrance rapidly and accurately in monocular video sequences. In order to minimize computation costs and overcome various complex scenes, based on image processing techniques a hybrid algorithm which combines motion analysis and model-based approach is applied. The proposed method with Intel Pentium IV operates at an aver age rate of 8 frames a second on the real world scenes where up to 15 persons come into the view of a single approximately vertically mounted camera.