Real-Time Traffic Sign Recognition Using SURF Descriptor

摘要

For road safety, traffic sign is essential for drivers by giving valuable safety and navigation information, pedestrians and even for the development of autonomous driver assistance system. Traffic sign can be classified by two methods can be approached. First approach is color base segmentation which is the region of traffic sign by using HSV color space (Hue, Saturation and Value) and the next approach is shape base segmentation using Hough Circle Detection. In thissystem, we use circle shape base detection for every traffic sign. At first, keypoints descriptor is extracted from each standard traffic sign image in database and then keypoints descriptor is taken from region of traffic sign extracted from Hough circle detection. After that, the nearest distance is matched keypoints descriptor between in each standard traffic sign image in database and extracted traffic sign image.