A novel vision-based lane detection and tracking algorithm using B-snake

Abstract

This thesis presents a research work on LDTUB (Lane Detection and Tracking Using B-snake) algorithm for AGV (Autonomous Guided Vehicle). Vision-based lane detection method is chosen here as a vehicle guidance strategy. There are several advantages regarding to this selection: (a) taking advantage of the existing visual cues, such as lane markings and road boundaries, which have been employed by our humans when controlling vehicles; (b) avoiding the modification of the road infrastructure. Other lane detection method such as cooperative approach, requires to add active or passive beacon to the road infrastructure.