A Research On Technology Of Recognizing Characters In Vehicle License Plates

Intelligent transport system is the trend of the development in national transport in 21st century. With the continuous developments of fundamental construction in traffic transport and vehicle administration system, the application of intelligent transport system based on the vision supervisor has been supplied with t he excellent basis. Of the i ntelligent transport system, license plate automatic recognition system is the very important aspect.License plate automatic recognition system can be divided five modules that are image acquisition, image processing, license plate orientation, license plate segmentation, character recognition. Although there are correlations in each module, they have respectively traits and technologies. In this paper, the research on character recognition in vehicle license plates is divided two sections. One is character recognition without training sets, the other is character recognition with training sets. In each section, we mainly research on the technology of character recognition from the feature extraction, template construction, matching formula improvement, etc.Under the circumstances of being no training sets, we use the theory of rough sets to construct grey-level matching templates based on the standard characters in vehicle license plate and effectively resolve some defects of bi-level matching templates. We represent the ways of automatically constructing the gray-level templates. Using the improved matching formula to recognize characters, we get better performances.Under the circumstances of being training sets, we use the rough reseau to extract features of characters. Taking advantage of the attribute reduction of rough sets, we efficiently delete the redundant features, reduce the number of the features and improve the efficiency of recognizing. New algorithm of attribute reduction and strategy of recognizing characters based on the influential factors are represented. Better accurate rate is reached.Under the circumstances of being training sets, we make use of the...