Positioning in a Tunnel Environment by Detecting Emergency Telephone Indicators Using Infrared Cameras

Abstract

Vehicle positioning is of great importance to driving assistance and vehicles’ navigation. In tunnel environment, positioning is challenging, since neither GPS sensors nor ordinary cameras function properly. In this paper, we propose a real-time solution for positioning of vehicles. In tunnels, we can gain the absolute coordinates of some specific objects, here the emergency telephone indicators, which are some special lights. Then during the travelling in tunnels, we sense the environment using infrared cameras and detect the emergency telephone indicators to infer the experimental vehicle’s position. In our detection method, we make use of both the emergency telephone indicators’ appearance information and motion information. Through well balancing the pipeline, by making time-consuming steps deal with fewer instances, we gain a real-time detection method with promising detection results.