License Plate Character Based On Discrete Hopfield Neural Networks Recognition Technology

In recent years, the number of China’s car with the development of economy has risen sharply, it brought great pressure to vehicle management department, the artificial registration license plate has failed to keep pace with the development of The Times. So the importance of automatic license plate recognition technology is obviously. License Plate Recognition technology is one of the important research topics in pattern recognition technology and computer vision in the field of intelligent transportation applications, it plays an key role in traffic management, used in many equipments such as Electronic police,parking lot management system, automatic management system of highway speeding, district vehicle management system and so on.This article in view of the license plate recognition in the process of license plate image preprocessing, license plate orientation, license plate character segmentation and license plate character recognition, this paper analyzed the key link of the study, and the main research is license plate character recognition. By using artificial neural network combined with digital image processing technology, complete a series of operations in the process of license plate recognition. By the prior knowledge, If the sample quantity is little when the BP neural network is running, the effect is not good. So it need a large number of test samples, which need longer time to complete a large number of calculations, license plate recognition system is not in conformity with the requirement for real-time by this way. After in combination with prior knowledge and repeated experiments in this paper, I decided to try based on the combination of edge detection and color feature extraction of license plate localization algorithm for license plate location; The vertical projection algorithm for license plate characters based on improved cutting; And try to adopt discrete the Hopfield neural network to combining with template matching algorithm, to complete the license plate character recognition. The biggest characteristic of discrete hopfield neural network is the associative memory function, which can not only to associative memory of incomplete character, it also can denoising which has been affected by the noise pollution.This is also the main basis of selecting the algorithm.In the process of research license plate recognition technology, after trial and error, I have designed a set of complete license plate recognition program, and use the Matlab to simulation. Practice shows that the way based on hopfield neural network identification technology can effectively accurately identify the license plate characters. I Believe that this technology has a high research significance and application potential in the development of intelligent transportation in the future.