SafeDrive: Enhancing Lane Appearance for Autonomous and Assisted Driving Under Limited Visibility (1807.11575v1)

Abstract: Autonomous detection of lane markers improves road safety, and purely visual tracking is desirable for widespread vehicle compatibility and reducing sensor intrusion, cost, and energy consumption. However, visual approaches are often ineffective because of a number of factors; e.g., occlusion, poor weather conditions, and paint wear-off. We present an approach to enhance lane marker appearance for assisted and autonomous driving, particularly under poor visibility. Our method, named SafeDrive, attempts to improve visual lane detection approaches in drastically degraded visual conditions. SafeDrive finds lane markers in alternate imagery of the road at the vehicle's location and reconstructs a sparse 3D model of the surroundings. By estimating the geometric relationship between this 3D model and the current view, the lane markers are projected onto the visual scene; any lane detection algorithm can be subsequently used to detect lanes in the resulting image. SafeDrive does not require additional sensors other than vision and location data. We demonstrate the effectiveness of our approach on a number of test cases obtained from actual driving data recorded in urban settings.