Following Dirt Roads at Night Time: Sensors and Features for Lane Recognition and Tracking

The robust perception of roads is a major prerequisite in many Advanced Driver Assistant Systems such as Lane Departure Warning and Lane Keeping Assistant Systems. While road detection at day-time is a well-known topic in literature, few publications provide a detailed description about handling the lack of day-light. In this paper we present multiple sensors and features for perceiving roads at day and night. The presented features are evaluated according to their quality for road detection. We generated a large number of labeled sample data and extracted the quality of the features from their probability distributions. The challenge of tracking an unmarked road under bad lighting conditions is demonstrated by comparing receiver operating characteristics (ROC) of the features at day and night-time. Based on these results we present a road tracking system capable of tracking unmarked roads of lower order regardless of illumination conditions. Practical tests prove the robustness up to unmarked dirt roads under different weather conditions.

[1]  Klaus C. J. Dietmayer,et al.  Night time road curvature estimation based on Convolutional Neural Networks , 2013, 2013 IEEE Intelligent Vehicles Symposium (IV).

[2]  Franz Kummert,et al.  Spatial ray features for real-time ego-lane extraction , 2012, 2012 15th International IEEE Conference on Intelligent Transportation Systems.

[3]  U. Franks,et al.  Lane Recognition on Country Roads , 2007, 2007 IEEE Intelligent Vehicles Symposium.

[4]  Michael Himmelsbach,et al.  Detection and tracking of road networks in rural terrain by fusing vision and LIDAR , 2011, 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[5]  M. Serfling,et al.  Road course estimation in a night vision application using a digital map, a camera sensor and a prototypical imaging radar system , 2008, 2008 IEEE Intelligent Vehicles Symposium.

[6]  Klaus C. J. Dietmayer,et al.  Augmenting night vision video images with longer distance road course information , 2013, 2013 IEEE Intelligent Vehicles Symposium (IV).

[7]  E. D. Dickmanns,et al.  A Curvature-based Scheme for Improving Road Vehicle Guidance by Computer Vision , 1987, Other Conferences.

[8]  Christoph Stiller,et al.  Non-parametric lane estimation in urban environments , 2014, 2014 IEEE Intelligent Vehicles Symposium Proceedings.

[9]  Cláudio Rosito Jung,et al.  Lane following and lane departure using a linear-parabolic model , 2005, Image Vis. Comput..

[10]  Uwe Franke,et al.  B-spline-based road model for 3d lane recognition , 2010, 13th International IEEE Conference on Intelligent Transportation Systems.

[11]  Giovani Bernardes Vitor,et al.  A probabilistic distribution approach for the classification of urban roads in complex environments , 2014 .

[12]  Michael Himmelsbach,et al.  Fusing LIDAR and Vision for Autonomous Dirt Road Following - Incorporating a Visual Feature into the Tentacles Approach , 2009, AMS.

[13]  Hans-Joachim Wünsche,et al.  Detection and tracking of rural crossroads combining vision and LiDAR measurements , 2014, 17th International IEEE Conference on Intelligent Transportation Systems (ITSC).