In order to realize an identification system of traffic road signs in real time, the authors have developed a performed detection method of these signs in real time. This method, presented in this paper, is implemented in two modules. The first is a pre-processing module based on video stream processing technics and the second, based on the polygonal approximation digital curves, identifies areas that may contain road signs using the particularity of their colors and contours. All algorithms implemented in this work are developed under the programming language C / C ++ using OpenCV library. The results of tests, realized on real videos of the traffic signs, show the performance improvement of the method developed in this work in terms of rate and speed of detection.
Interpolation implementation of Bayer pattern to PLIC in video stream processing tasks