The Image Processing Of Road Traffic Sign

With the development of economy and the improvement of people's living standard. There are more and more vehicles in our country, the ITS (Intelligent Traffic System) has gained more and more attention. As an importment part of the ITS, the RTS (Recognition of Traffic Signs) plays a more and more important role in road safety. So studying on road traffic signs recognition system has important theory and practical values.In this paper, we do some research on detection and recgnition of prohibition and warning traffic signs. After analyzing the prior information which is color and shape characteristics of traffic signs, we make detection and recognition of traffic signs. We present histogram equalization color image enhancement based on multi-channel, which make histogram equalization in each channel of RGB color space, I(intensity) channel and S(Saturation) channel of HIS color space. In this way, the uneven illumination and color distortion problems can be resolved. We use the threshold segmentation algorithm in RGB color space and in H (hue) channel of HIS color space respectively, and make some comparison between them.Secondly, we set a threshold to remove small disturbance areas, and calculate the boundary invariant moments. We compare them with standard templates of circular and triangular to position the region of traffic signs. This method can lower computational complexity than that in Hu's invariant moments.At last, we make a brief introduction of AIN (Artificial Immune Network). After extracting the Hu's invariant moments of the traffic signs, we present a new classification algorithm of artificial immune network to classify the traffic signs. The experiments indicate that the new method of traffic signs recognition and classification is quick and accurate.