Localization of people or objects is becoming an increasingly important task in a wide area of ITS-related applications. In this paper we address the problem of localizing people in public transport vehicles, e.g. as part of an automated ticketing system. In our application positioning information is primarily acquired by means of radio detection and ranging in a wireless sensor network (WSN). We show how to further improve the accuracy of the positioning estimate by data fusion with the vehicle's surveillance camera system. The possible role of cameras in the localization process is examined through the example of a single person that is observed from distinct viewpoints. We suggest a data fusion method that is based on combining 2D-Gaussians in the WSN's coordinate system and show how incorporating positioning information from cameras can improve the overall localization accuracy.
Camera-assisted passenger localization in public transport vehicles
2013
10 Seiten, Bilder, 8 Quellen
Aufsatz (Konferenz)
Englisch