Large Area Traffic Monitoring with Serial Images from Airborne Cameras

Abstract

In order to obtain precise knowledge about the traffic flow of a larger area, only airborne cameras can give an actual situation report of all streets covered. The presented method for determining several traffic parameters for single vehicles and vehicle groups involves recording and evaluating series of digital aerial images acquired from high altitude (2000 m to 3000 m) and with a large total field of view. The camera system used is a combination of 3 commercial digital cameras with 48 MPixel and a coverage of about 7 km x 2 km for a single image. Investigations involving visual and automatic image interpretation are described. The recording frequency of the individual images is preferably 3 Hz or more, especially for automatic tracking. The accuracy and possibilities of the technique are analyzed and presented. The traffic data obtained from induction loops or other local instruments can be checked with the help of this method and local traffic anomalies like congestions can be precisely identified and studied. The paper shows how the image data are acquired, processed and evaluated and how this methodology can be used for traffic monitoring.