Road Markings and Road Edges Mapping With Inverse Visual Detector Model

The paper proposes an algorithm for mapping linear features detected on the roadway – road markings, curbs, road boundaries. The algorithm is based on an occupancy grid mapping method with an inverse sensor model. We propose an inverse sensor model that takes into account the spatial error of the linear feature detector and study the influence of various parameters of the model on the resulting quality of mapping. The mapping algorithm was tested on data recorded on a real autonomous vehicle. The quality of the mapping algorithm was assessed by several common quality metrics. In addition, the mapping problem was considered as a binary classification problem, in which each map cell may or may not contain the linear feature, and the ROC curve was used to assess the quality. The experimental results have confirmed that the proposed algorithm can effectively filter detection noise and false detections, which confirms the applicability of the proposed algorithm and the inverse sensor model for solving practical road markings mapping problems.