Lane recognition based on location of raised pavement markers

Lane recognition plays an important role in driver assistance systems. Many Lane recognition methods have been proposed until now. But most methods are used to recognize the traditional painted lines and can´t work on the road where there are only raised pavement lane markers such as reflective markers and bots dots. Lane recognition on this kind of roads is a difficult problem especially in the daytime for the raised pavement markers are poorly visible and not reflected. In this paper, a novel solution for lane recognition in this case is presented. Firstly, the dot filters and the OLS templates are proposed for location of raised pavement markers. Then a line fitting method based on Hough transform and clustering is used on the accumulation image of raised pavement markers to detect the implicit lane. The experimental results under various scenes show that the proposed method is effective and can be implemented in real-time.