摘要 |
A method for analyzing road image sequences from the on-board camera of a moving vehicle. The method combines the analysis (130, 131, 132) of a luminance signal, from which outlines are determined, with the analysis (110) of colour video signals, while taking into account the fact that the colorimetric components of a road area in an image contain no red or green cast, in order reliably to locate the edges of the road in the image on the basis of said outlines. A road mask is then formed by filling. The method may be used in particular to detect obstacles on the road on which a vehicle is travelling, and the data generated may be combined with data from other detectors in a road traffic obstacle detection system. |