Traffic Warning Signs Detection and Recognition Based on Fuzzy Logic and Chain Code Analysis

This paper presents an algorithm for the detection and recognition of traffic signs in a category of warning sign. The algorithm could be divided into two phases: detection and recognition. Images were captured from a survey vehicle camera surveying actual road environment. The detection step used HSV color model to process and detect the region of interest using color of the traffic sign by fuzzy logic. The recognition step processed the traffic sign symbols and transformed them into chain codes which were compared with the chain codes in the database by using longest common subsequence algorithm. The algorithm was applied to more than 4000 images and produced an accuracy over 90% with average errors of 8.9% positive false and 9.71% negative false.