A Study on Vision Based Lane Detection Methods for Advanced Driver Assistance Systems

Abstract

Intelligent driving systems need to find the lines that show where the lanes are present. It can help drivers prevent lane hopping and improve vehicle positioning and identification by providing information about the current road conditions. The lane detection encounters several obstacles, including harsh illumination conditions, missing lane markings, and impediments. Due to their outstanding performance, Artificial Intelligence, machine learning, and deep learning-based algorithms have recently attracted considerable interest in the intelligent driving society. In this paper, we thoroughly analyses different lane detection approaches for lane detection including deep learning based techniques. In addition, we review known datasets about lanes and assessment criteria. It ends with a discussion of current problems and possible directions for a lane detection system.