A method for road markings extraction on urban streets acquired by a camera mounted on a moving vehicle is described. This method works in three stages. First, the search for the road markings is reduced to a suitable bird's eye view of road surface by using inverse perspective mapping. Secondly, an integrated approach for image segmentation is presented that combines local adaptive threshold and canny edge detection, which output a binary image by separating desirable foreground objects from the background. Thirdly, a geometrical analysis of the contours extracted from binary image is carried out, which extract candidates for further road markings recognition. This method can be helpful for road markings detection under uneven illumination conditions.