Options
1998
Book Article
Titel
Robuste Fahrspurverfolgung und Bewegungskompensation in einem videobasierten Echtzeitsystem
Alternative
Robust lane tracking and motion compensation in a video-based real-time system
Abstract
We present fast and robust video-based components for real-time driver assistance system. The system can detect, track and classify several marked or unmarked lane borders. Furthermore, an obstacle detection module has been integrated. Various traffic situations, road types and changes in illumination require great care on robustness and reliability. This was achieved by evaluation of image sequences taken from a high dynamic range camera as well as by applying reliability tests of the results within each system component. Moreover, the results of the obstacle module are used in the lane tracking module and vice versa. This results in an increased overall system performance. The system has been integrated into some experimental vehicles. Experiments have been performed online in the cars as well as offline with more than 150. 000 images taken under a variety of environmental conditions. The system performed well under different traffic situations and weather conditions.