The Research Of License Plate Segmentation And Recognition System Based On Fuzzy Theory

License plate recognition(LPR) is an important application that calculator sense of vision and mode identify realm at the intelligence transportation, is one of the core techniques which carries out an intelligence transportation system. This paper analyzes the international development condition and trend of LPR firstly, then tallies up currently LPR a little bit shortage and places that need to be improved, puts forward a research project of license plate segmentation and recognition system based on fuzzy theory.Firstly, according to deeply research in fuzzy threshold segmentation in image segmentation, to establish a research program of this paper Then, to put forward an improved license plate image segmentation algorithm based on fuzzy threshold and prove its superiority through the experiment of a great deal of license plates image segmentation. Finally, this paper presents a method to extract the feature of characters based on coarse grid feature and projection feature, and builds up a fuzzy neural network recognition system which is based on BP neural network to carry on character recognition.Through analyzing the outcome of license plate character recognition and various experimental data, this algorithm is proved possible and superior, and the recognition rate of Chinese character is over 96%, and the number and letter characters recognition rate is beyond 98%.