This paper presents an algorithm for road lane recognition. The algorithm composes of the estimation of 1D function of the histogram of gradient magnitude with respect to the image gradient direction by a recursive filter, determination of a band regi...
This paper presents an algorithm for road lane recognition. The algorithm composes of the estimation of 1D function of the histogram of gradient magnitude with respect to the image gradient direction by a recursive filter, determination of a band region as the search region for the road lane, construction of 10 function of the histogram of gradient magnitude of edge pixels in the band region, and recognition of road lane by the convolution of two functions. We also initially estimate the vanishing point and the location of road lane for host vehicle based on calibrated camera parameters. To prevent the failure of road lane: detection due to the disappearing of a lane mark and the lateral motion of host vehicle, this algorithm utilizes the recursive ftltering. The newly proposed approach provides the reliable detection of road lanes of natural outdoor scenes even in the presence of various noises.