Research Of License Plate System Based On SVM

License plate recognition system is an important part of ITS(Intelligent Transportation System),which has a wide range of applications.A license plate recognition system mainly consists of three parts: license plate location,character segmentation and character recognition.This thesis focuses on the research of character segmentation and character recognition.In the aspect of license plate positioning,this thesis first analyzes the characteristics of China's automobile license plate positioning,then adopts the vehicle license plate positioning method based on mathematical morphology,then adopts the improved Bernsen algorithm to binarize license plate image,and conducts experiments on the sampled data.In the aspect of character segmentation,this thesis first determines the upper and lower boundaries of the license plate character area according to the license plate specifications,then uses the character segmentation method based on the second character,and finally adjusts the width of the first character and the last character.The experimental results show that this method has a fast cutting speed,and ensures the accuracy of character segmentation.In the aspect of character recognition,this thesis designed four different SVM classifiers of Chinese characters,letters,letters or Numbers and Numbers according to the arrangement features of license plate characters.In terms of character feature extraction,the performance of classifier is compared when 32-dimensional rough grid feature,128-dimensional rough grid feature and 192-dimensional peripheral contour feature are extracted.In the aspect of SVM multi-classification,this paper studies and compares two multi-classification methods of "one-to-one" and "one-to-many".Through experiments on the collected sample data,the recognition rate of the multi-classification method of "one-to-many" is higher than that of "one-to-one" multi-classification method.In this paper,the traditional "one-to-many" multi-class classification method is improved and the sample data is tested.In this paper,a new method of recognizing similar characters is proposed to solve the problem of confusing similar characters.