CHARACTER RECOGNITION OF LICENSE PLATE IMAGE BASED ON MULTIPLE CLASSIFIERS

An approach of character recognition of license plate image based on multiple classifiers is proposed in this paper. For numbers and English characters, features are extracted from binary character images by ET1, DT12 and jumping. For Chinese characters, features are extracted from gray-scale character images by Gabor filters which are specially designed from statistical information. In order to access a high recognition rate, 3 classifiers arc used. They are SVM, BP ANN and minimum distance classifier. All the classifiers are arranged by a structure of voting. Experiments show that the proposed method has effective performance on Chinese character image recognition.