Development and Implementation of Lane Departure Warning System on ADAS Alpha Board

Abstract

The increase in vehicle safety and autonomy is achieved by the implementation of Advanced Driver Assistance Systems (ADAS). ADAS consists of an appropriate computer platform that processes information collected using various types of sensors during driving. Usually, ADAS warns the driver or acts independently on the vehicle control system (e.g. automatic braking) to prevent possible accident. One of the very common ADAS is a lane departure warning system which warns the driver when a vehicle starts to leave a driving lane unintentionally. Such system is based on the processing of an image obtained by a front-view camera. In this paper one such system is proposed which is based on traditional computer vision techniques and is implemented on ADAS Alpha board with Texas Instruments TDA2xx System-On-Chip. The proposed system is evaluated on images taken from Kitti dataset and on recorded video sequences taken in suburban areas.