Design and FPGA implementation of dual-stage lane detection, based on Hough transform and localized stripe features

Abstract

Robust, accurate and real-time detection of lane boundaries based on embedded hardware is an essential element of several driver assistant systems. In this work we present an FPGA-based dual-stage lane detection algorithm to cope with real world challenges such as cast shadows, occlusion of lane markers, brightness variations, wear, etc. In first stage, Sobel operator and adaptive threshold are used to extract lane edges, followed by Hough transform to extract the road markers. Second stage of the algorithm operates on original grayscale image and identifies stripe features near several candidate points with highest probabilities to find the landmarks. These extracted features are then used to detect the lane boundaries with high accuracy. Experimental results based on FPGA platform under various road conditions obtained from various datasets indicate that our algorithm can process about 60 frames per second for 720 pixels video input. Lane detection accuracy of 94.3% is achieved in average which may reach up to 97.8% in low congested highway during daylight.