A Robust Lane Detection and Verification Method for Intelligent Vehicles

Abstract

Robust lane detection is important to the lane departure warning (LDW) for driver assistant system. In this study, lane marks are extracted by searching the lane model parameters in a special defined parameter space without thresholding. The proposed method is based on the lateral inhibition property of human vision system to clear up the edges of lane marks in variant weather conditions; moreover, the lane detected results are verified by the proposed conjugate Gaussian model such that there are no false alarm on the edges of shadow and other vehicles. The proposed lane detection method can gain precise lane-mark information for a lane departure warning system.