An Effective Method for Traffic Signs Segmentation

An effective method for traffic sign segmentation is proposed. Through the selection of appropriate characteristic operator on the basis of the distinctive color features of traffic sign, it first quickly obtains gray image of chromatic aberration. Then, the Otsu´s thresholding algorithm is subsequently applied to locate accurately the candidate regions of the traffic sign. The experimental result shows that the proposed algorithm can extract the traffic sign from the background well up to the technical standards under various natural illuminating conditions, and the segmentation effect is superior to that performed by the generic segmentation with steady threshold, and the performance in shape checking is thus improved. The presented approach is featured in good robustness, high speed, and as a result can be potentially applied to the real-time processing and commercialization.