Traffic Road Sign Detection and Recognition in Natural Environment Using RGB Color Model

Traffic sign detection and recognition play crucial roles on the Intelligent Transportation System. So far, color-based traffic sign detection and segmentation have been widely used for feature extraction and detection. This paper presents an analysis of the performance of five different color models for the color segmentation and subsequent detection of traffic signs in two-dimensional static images that obtained in real-world environment. Firstly, using color thresholding techniques to isolate relevant color region (red, blue) from the image. The regional morphology processing algorithms is applied in order to extract traffic sign's region of interesting (ROI), it could remove the noise and isolate the traffic sign. Then, a rectangle region in the original image to be selected according as its shape property. Finally, a way of quantitatively evaluate the performance of the different color space detection algorithm on the widely-used German Traffic Sign Detection Benchmark (GTSDB) has been proposed.