Fahrstreifenerkennung mit Deep Learning für automatisierte Fahrfunktionen.

This contribution is concerned with the estimation of the road topology and detection of driving lanes based on images of a front-facing camera. Herein, the road topology represents the number of drivable lanes within a certain range ahead of the vehicle. The proposed solution uses Convolutional Neural Networks and the learned features are examined. Finally, the experimental results and future research directions are discussed.