Research On The Technologies Of Vehicle License Plates Recognition

The number of cars has increased dramatically in our country because of the high speed development of economy and the important means of transportation. This situation led to a higher level to traffic and security management. The rhythm of social life is accelerating in recent years. The vehicle management is also turning its mode from the traditional manual ways to the intelligence and automation gradually. The recognition of car license plates for vehicle management provides a good visual conditions. The license plates recognition technology can improve the standard of transportation system in our country. It also can help us to meet the demands of intelligent transportation. Therefore, the research on the key technology of vehicle license plates recognition has its important significance to improve the reliability of whole recognition system. The main study in this paper is as follows:1) Research on several key image processing technologies which have been used in car license plates recognition. Doing research on the grayscale enhancement, histogram transformation and some other technologies which is related to license plates image information. Then filter out the noise from the image and detect the edge. It can effectively filters out noise and the weakly correlated information from the original images. It can also enhance the strongly correlated information. After this, the license plates image preprocessing has been finished.2) Research on the main component of car license plates automatic recognition. Firstly, mainly analyse the license plates location method which based on mathematical morphology. Make a judgement to the candidate regions appeared in our processing. Then proposed a method combined with the feature of image color judgement. And then use the vertical projection to do the further processing and complete the license plates character segmentation. Analyse and propose solutions with the abnormal state which may be caused by the character structure. Finally, complete the license plates character recognition by using the template matching algorithm.