License plate character recognition method based on combination feature and BP network

Abstract

In order to improve the license plate character recognition rate, a license plate character recognition method based on combination feature and BP neural network is proposed. Firstly, according to the license plate character texture features, the basic LBP operator is improved in our method. Secondly, the improved local binary model and horizontal vertical projection are combined to extract the characteristics of the license plate character image. Then the combined feature is used to train the classifier in BP neural network, and it is applied to identify license plate characters. The experimental results show that the recognition accuracy rate of the license plate reaches 94 .55% . The validity and robustness of the method are verified.