The Recognition of the Vehicle License Plate Based on Modular Networks

This paper presents a novel method of real-time recognizing the vehicle license based on the integrated modularized neural network. Some locations in the license plate contain Arabic numeral only and the others contain Arabic numeral, English letter or Chinese character. According to the complexity, the two kinds of network structures, ω i /ω j and ω i /ω i –, are organized and the outputs are integrated under the given fuzzy rules to perform the recognition tasks. The results of the experiments have shown that this method can recognize the characters quickly and accurately.