Traffic sign recognition based color model

Abstract

The main problems that encounter the traffic light detection algorithm have become a handicap to the performance of the algorithms. Problems associated with the change of sign color due to bad weather and illumination changes of sunlight make the detection hard task. In the current work, we discuss these problems and propose a new idea of an efficient real time color sign recognition that relies only on color information. The proposed approach is based on building a red-model in hypothetical red, green, blue (RGB) cube using a large database of traffic signs. The segmentation has been implemented on the traffic signs that hold red color only as an example to illustrate the proposed approach. Results showed that the proposed algorithm is accurate as well as the computational cost is reduced.