A novel algorithm for lane detection based on iterative tree search

An important module in driver assistance systems is the lane departure warning which warns the driver when the vehicle begins to cross out of its lane inadvertently. A novel lane detection algorithm designed for lane departure warning system is presented in this paper. The proposed approach exploits the advantages offered by tree search to set up an intuitive and lightweight lane detection solution compared to existing ones. The algorithm is capable to work with poor light conditions, e.g., tunnel road, irrelevant markings on the road and the inherent long and thin property of lanes, scenarios with night conditions. Furthermore, the algorithm has a low time complexity which allows to be used on small and cheap hardware systems. An experimental setup containing a low cost front-facing camera sensor and an embedded computing board are used to implement the algorithm. A testing in a real highway scenario is carried out to validate the effectiveness of the proposed algorithm.