Traffic sign recognition by fuzzy sets

A novel fuzzy approach developed to recognize traffic signs is presented in this paper. More than 3400 images of traffic signs were collected in different light conditions by a digital camera mounted in a car and used for developing and testing this approach. Every RGB image was converted into HSV color space and segmented by using a set of fuzzy rules depending on the hue and saturation values of each pixel. Objects in each segmented image are labeled and tested for the presence of probable sign. Objects passed this test are recognized by a fuzzy shape recognizer which invokes another set of fuzzy rules. These fuzzy rules are based on four invariant shape measures which are invoked to decide the shape of the sign; rectangularity, triangularity, ellipticity, and the new shape measure octagonality. The method is tested in different environmental conditions and it shows high robustness.