Traffic Sign Detection and Recognition Using Feature Based and OCR Method

Abstract

Traffic sign detection is a technology by which a vehicle is able to recognize the different traffic signs located on the road and used to regulate the traffic.
Traffic signs are detected by analyzes color information contained on the images having capability of detection and recognition of traffic signs even with bad visual artifacts those originate from different conditions. The feature based method is intended for traffic sign  detection, in this method two sets of features are to be detected in the reference and sensed images, identifying key points in the images and match  among those points to find the similarity, the SURF descriptor is used for key points and point matching. After detecting the shape of the traffic sign the optical character recognition (OCR) method is used to recognize the character present in the detected shape. A technique, based on Maximally Stable Extremal Regions (MSER) region and canny edge detector is supervised for character recognition in traffic sign detection.