Multiple lane boundary detection using a combination of low-level image features

In this paper, we propose a vision-based multiple lane boundaries detection and estimation structure that fuses the edge features and the high intensity features. Our approach utilizes a camera as the only input sensor. The application of Kalman filter for information fusion and tracking significantly improves the reliability and robustness of our system. We test our system on roads with different driving scenarios, including day, night, heavy traffic, rain, confusing textures and shadows. The feasibility of our approach is demonstrated by quantitative evaluation using manually labeled video clips.