Detection, Tracking and Classification of Road Signs in Adverse Conditions

In this paper a complete automatic system for road-sign detection, tracking, and classification is presented and evaluated. The processing of video frames in the L*a*b color space improves significantly the sign detection rate by processing the same frame in different normalized color spaces. The tracking module reduces significantly the processing time by transferring the sign detection information in the next frames and processing different radii signs in parallel. The proposed system is evaluated in normal, raining and night driving conditions. In a total number of 266 road-sign recordings, the complete system track and recognize successfully 216. The main source of system fault appears in city night driving due to the presence of a great number of light sources