The Research And Method Of License Plate Localization And Character Recognition

License plate recognition system plays an important role in many applications,such as parking system,road monitoring,highway toll collection system,electronic police system.In our country,there are many kinds of license plate,and it has numbers,letters and Chinese characters.In this paper,an automatic license plate recognition system based on artificial neural network is proposed.The license plate recognition system consists of three main components: license plate location,character segmentation and character recognition.(1)In the license plate positioning process,it considers mainly color characteristics of the plate and texture features.Firstly,locate the license plate accord to the color couple which the background of license plate and characters,and then consider the texture features of license plate to identify the upper and lower edge,cut out the license plate area.This method does not depend on the shape of the license plate,so the deformation or bending of the plate is still valid.(2)The character segmentation of the license,the first to cut image pre-processing,and then use the Otsu method of license plate image binarization.After the work,according to the image of the black and white jump times to remove the border and rivet,and then set the threshold value according to the character width to carry out the segmentation character.(3)The artificial neural network is used in the research of license plate recognition.Training network requires a large number of sample data.13 feature extraction method is used to extract the feature of the recognition character,to collect the sample data.Then the sample data is loaded into the neural network for training.In the recognition of characters,numbers,letters and Chinese characters are separated,the highest recognition rate is 95%,the time is less than 1s.