Driver-assistance systems., which monitor driver intent, warn drivers of lane departures, or assist in vehicle guidance, are being actively considered. Guardrail detection is one of the most important systems. In this paper, an efficient approach is presented for reliably detecting road guardrail using camera. The proposed approach has the capability of locating and detecting guardrail accurately on both straight and curved roads. Firstly., a top-view of the road using inverse perspective mapping (IPM) and a simple parabolic guardrail model are presented. Secondly, a bank of steerable filters is applied to ensure the matching templates. Finally, an algorithm for detecting guardrail using the matching rule is proposed in this paper. The experimental results show that the system has robustness to environment, and the fully implemented system runs at near-real-time speeds of 21 frames per second.
展开▼