Milepost sign detection

In this paper the problem of the detection of milepost signs, a specific type of road sign showing distances, placed on Italian highway has been faced. First step concerns the robust identification of rectangular sign by optical flow analysis. Second step concerns rectangular sign detection based on searching grey level discontinuity on image and Hough transform. Classification is based on analysis of surface color on inner part, detection of shape and color of rectangular border around the sign, characters analysis. Our approach can detect sign candidates in presence of complex background.