License Plate Recognition Technology Based on Neural Network

Abstract

License plate recognition is an important part of intelligent transportation system. It plays an irreplaceable role in solving the problem of vehicle license plate issuance and improving the level of urban road traffic management. In this paper, the method of vehicle location and character segmentation is studied by using artificial NN algorithm. First, the BP Neural Network (NN) training method is used to complete the fusion model from the image extraction feature parameter set to the sample classification to the input mode selection and output mapping, the calculation and analysis of the rotation distance of the license plate area, and the recognition of the background color of the license plate. Then, the correctness and feasibility of this strategy are verified by the MATLAB simulation experiment results. The test results show that the learning time of the license plate number recognition technology based on the NN algorithm is between 20s and 30s, The recognition time is between 12s and 16s, and the correct recognition rate is above 90%. This shows that the algorithm is excellent in optimizing the license plate recognition technology.