Road Speed Sign Recognition Using Edge-voting Principle and Learning Vector Quantization Network

Abstract

This paper presents an automatic speed sign detection and recognition for providing the visual drivingassistance of speed limits awareness. To reduce the influence of digital noise caused by lighting condition and pollution, a segmentation based on pan-red color information is applied to extract the shape of speed sign. Based on the edge-phase information of a circle shape, a novel edge-voting principle is proposed for fast detecting the speed sign candidate from road scenes. The recognition of the content of speed sign is achieved through a modified learning vector quantization (LVQ) network which also verifies each candidate to eliminate nontarget blobs. Results show a high success rate and a low amount of false positives in both detection and recognition strategy under a wide variety of visual conditions.