Reliable lane detection is a key component of autonomous vehicles supporting navigation in urban environments. This paper introduces the GOLDIE(Geometric Overture for Lane Detection by Intersections Entirety) system, a vision-based software architecture that uses an on-board single camera to determine the position of road lanes with respect to the vehicle. We propose an efficient vision-based lane-detection system that combines an appearance-based analysis with salient point tracking. The appearance-based analysis consists of segmenting high contrast areas that fit inside a Region-Of-Interest(ROI) on the frame. The salient point tracker selects interesting points based in a reference line, that guides a dynamic ROI. The tracking ROI look for paint lane marks close to the last lane reference found, where road marks are likely to emerge, in order to maintain the usability of the salient point tracker. The tracking is performed with the Lucas-Kanade algorithm and the lane points candidates are selected according to a predefined triangular model. Once such lanes points are detected, the vehicle position is estimated based on the intersection of linearised lanes determined through a vanishing point approach. Experiments and comparisons with other algorithms illustrate the applicability of the method.

The lane detection is a vital component of autonomous vehicle systems. Although many different approaches have been proposed in the literature it is still a challenge to correctly identify road lane marks under abrupt light variations. In this work a vision-based ego-lane detection system is proposed with the capability of automatically adapting to abrupt lighting changes. The proposed method automatically adjusts the feature extraction and salient point tracking cues introduced by the GOLDIE (Geometric Overture for Lane Detection by Intersections Entirety) algorithm. The variance of the lighting conditions is measured using hue-saturation histogram and abrupt light changes on the road are detected based on the difference between histograms. Experimental comparison with previously proposed algorithms demonstrated that this method achieved efficient lane detection in the presence of shadows and headlights. In particular, the accuracy of the algorithm applied on the footage with highest light variation increased 12.5% on average. The overall detection rate increased 4%, which illustrated the applicability of the method.

No abstract