In this work, we propose a light method for the recognition of traffic signs. This new approach involves three steps which are the recognition of the shape of the border, the border color, and recognition of the pictogram. To extract the color, the RGB color representation is used. The method of profile projection (outermost point method) is used to extract the feature vector associated with the border of the sign and the vector associated to the pictogram which is located inside the border. A reference base for borders as for pictograms is defined and used as a tool for decision to identify a traffic sign. The results are quite conclusive, both the recognition and speed.