Improvement of crowd management through an automatic queue length measurement
AUBERT
Type de document
COMMUNICATION AVEC ACTES INTERNATIONAL (ACTI)
Langue
anglais
Auteur
AUBERT
Résumé / Abstract
In the framework of the European project CROMATICA, some automatic video analysis systems were developed to improve the public transport quality of service. These systems encompass various incident detections ('abnormal stationarities', 'counter-flow', or 'people falling on the track', for instance). The current article will pay attention to an automatic measurement of the passengers queue length. The system presented is robust to contrast changes and it is able to detect events even under low contrast. It is based on the representation of the images by their level lines, and on a study of the effects of contrast changes on them. This algorithm has been tested on real airport situations.