PathMark: A Novel Fast Lane Detection Algorithm for Embedded Systems

摘要

Lane detection is an important problem in the field of intelligent transportation and safety driving, for which computer vision approaches are widely used. However, previous works involve unbalanced performance between robustness and speed, which makes it hard to be practical. In order to meet the requirements of safety driving, we present a new, fast and robust lane detection algorithm. It is based on searching lane marker candidates by the brightness feature and geometric matching from discrete segments to full lane markers, and then followed by a random noise filter. PathMark has been validated on a wide range of real world driving sequences, showing capability of extracting all kinds of lanes from different road conditions. Compared with previous work, PathMark operates at doubled frame rate on the embedded platform with comparable detection rate. PathMark is more suitable for real safety driving use in embedded systems.