LDWS for Smart Vehicles Using Image Processing Based On Hough Transform

Abstract: This paper proposes a method for monitoring the position of a vehicle on a lane. Cameras are mounted on the sides of vehicle and focus downwards on the road to capture images containing the lane lines and wheels/vehicle boundaries. The Hough transform (HT) is employed to detect the lane from the captured images. The correlation of sequential images and the constraints of lane lines and cameras are used to minimize the region of interest (ROI) in both image space and HT space, which increases the capture accuracy and decrease the computation complexity. The lane position monitoring system is assumed to be useful for unmanned vehicles, and can also provide warnings in normal vehicles to enhance driving safety. The experiments show the method is effective on various roads and driving situations.

Keywords: Hough Transform; Lane line detection; Video content analysis; Unmanned vehicle