Real Time Multi-lane Detection Using Relevant Lines Based on Line Labeling Method

We propose a real time multi-lane detection algorithm using monocular camera mounted in front of the vehicle. The proposed algorithm is focused on reducing computation time by using line labeling method. Through the line labeling method, the algorithms requiring high computation time such as RANSAC and vanishing point estimation are able to process efficiently. The proposed algorithm is evaluated on both our dataset and Caltech Lane Dataset. The experimental results show that the computation time of the proposed algorithm takes 18.7ms per frame in our dataset which is suitable for real time application and achieving comparable accuracy to previous study.