Hough Transform-Based Road Detection for Advanced Driver Assistance Systems

摘要

This paper proposes a real-time road region detection method for safe driving assistance systems. This is a method on detecting road regions by receiving input images from vehicle black boxes. As such, we propose a method that detects major straight line components through Hough transform in input images, selects a vanishing point from the intersection of detected straight lines using preliminary information from the road environment, and detects road regions using the selected vanishing point and left and right straight lines from the Hough transform. As a result of applying the proposed method to various road environments, approximately 0.37 s were consumed per frame, providing over 81% detection accuracy.