Robust Road Sign Segmentation

This paper deals with the robust segmentation of road signs under bad illumination conditions including night-time. The road sign recognition system is very important for safe driving. The proposed segmentation system is developed using relative similarity with the two-way connectivity. Here, we focus mainly on how to obtain the optimum threshold to each illumination condition. The proposed system works on various types of road signs such as circular, equilateral triangle, pentagon and diamond-shapes. The final goal of our proposed system indicates to the driver the presence of a sign in advance so that some incorrect human decisions may be avoided. The effectiveness of the proposed method is demonstrated using 100 images of road signs under various illumination conditions. The result is encouraging and demonstrates the ability and efficiency of our algorithm in achieving the segmentation task. According to our experimental results, the average accuracy rate for segmentation is 98%. 1