Vision-based lane detection for an autonomous ground vehicle: A comparative field test

Abstract

We examine the problem of designing computer vision algorithms to autonomously drive an off road vehicle between two lane markings painted on the ground. In this paper we describe field tests used to compare the efficacy of two popular line extractions techniques from the literature: the Hough Transform and the RANSAC Algorithm. Although it is very implementation dependent, we found the Hough Transform to be superior to the RANSAC algorithm in both speed and accuracy for identifying lane markings in the off road environment.