A computer vision system for measurement of pedestrian volume

This study applies image sequence analysis techniques to perform pedestrian volume measurement. A new algorithm, which consists of four steps, was proposed. The algorithm was capable of extracting pedestrian information from consecutive video images while eliminating vehicle information. The volume of the pedestrians crossing a detection area can be automatically derived by the algorithm. A computer vision system, called the Pedestrian Volume Measurement (PVM), was developed based on the algorithm. The significant advantage of the PVM system to previous systems was that the new system can be used at intersections where vehicle presence makes the previous systems not practicable.<>