An online traffic sign recognition system for intelligent driver assistance

This paper proposes a system that is capable of detecting, recognizing and indicating traffic signs to the driver to provide a safe and convenient driving experience. The system is composed of three phases; image pre-processing phase, detection phase and recognition phase. Image pre-processing phase deals with image segmentation that is employed to enhance the images to obtain clearer information under variable lighting conditions. During the detection phase, the objective is to find traffic signs in the input image frames and to return the regional coordinates of detected signs. In order to detect traffic signs, a cascade classifier was trained to analyse the frame and to detect signs in each frame. In the recognition phase, a multiclass support vector machine image category classifier was trained to recognize and alert the driver. The system comprised of these phases is capable of detecting, recognizing and alerting traffic signs with 98.6% accuracy while the vehicle is moving at a speed in the range of 40-45 km/h.