An illumination-invariant nonparametric model for urban road detection using monocular camera and single-line lidar

In this paper, we propose an illumination-invariant nonparametric model for urban road detection based on a monocular camera and a single-line lidar. With the monocular camera, we can accurately locate road vanishing point after removing the adverse shadowy effect by an illumination-invariant image representation. With the constraint of detected vanishing point, we propose a Dijkstra method to compute a minimum-cost map, where the minimum-cost path from the vanishing point to any other pixel can be found. With the single-line lidar, we can locate two key points in the image bottom that correspond to two road-border points. Thereafter, the two road borders can be found as the minimum-cost paths that originate from the vanishing point to the two key points, respectively. The proposed method has been tested on over 4000 images of the KITTI-Odometry Dataset [3] and the Oxford Robotcar Dataset [7]. Experimental results demonstrate that the method achieves promising performance on a variety of road scenes.