Intelligent intersection : artificial intelligence and computer vision techniques for automatic incident detection


Type de document
Résumé / Abstract
Our contribution presents the first integration and evaluation tests of the intelligent intersection system. This system is based on several promising technologies such as object-oriented design, real-time knowledge-based systems and a new generation of sensors (video sensors). First we succinctly give some formal specifications based on an object approach of the models used, representing both static and dynamic aspects of the real world of the intersection. Then the experimenting process of the intelligent intersection system is presented and particularly the field trial work program partly achieved of two drive projects concerned by the automatic incident detection function. Three pilot sites are involved in this project: Valencia in Spain (invaid I and II), INRETS site in France (invaid I and II) and Lyon in France (LLAMD-LEADER). At present, on the basis of the first experiments we have obtained technical and operational interesting results. However, these results can be obtained only by the implementation of a specialization of the model used and an adequate calibration of the video measurements. These remarks led us to reflect on the methods for analysis and design to cover the genericity gaps of the model and to relaunch research works on the measurements used.

puce  Accès à la notice sur le portail documentaire de l'IFSTTAR

  Liste complète des notices publiques de l'IFSTTAR