The License Plate Recognition System Based on Fuzzy Theory and BP Neural Network

Abstract

In different conditions such as light and complex backgrounds, we get some car images, the traditional methods are slow convergence speed and low accuracy. This paper presents a method which applies fuzzy theory to enhance several features of for target. To obtain the license information, we use an improved BP neural network algorithm, by through setting proper numbers of hidden layer of BP network, we can solve the recognition problems of China's automobile license such as characters kinds, the numbers, and confusing. This method can improve the accuracy and efficiency of car license recognition, and enhance the system robustness.