基于纹理特征与ν-SVM的车牌汉字识别  

Recognition of Chinese Character on License Plate Based on Texture Features and v-SVM

作　　者：束赟[1] 

机构地区：[1]东南大学学习科学研究中心,南京210096

出　　处：《信息化研究》2012年第2期38-40,54,共4页INFORMATIZATION RESEARCH

摘　　要：车牌汉字识别是车牌字符识别中的难点,文章通过对二维Gabor小波特征以及支持向量机(Support vector machine,SVM)分类器进行详细的分析,采用二维奇对称Gabor小波作字符特征滤波函数,并通过基于ν-SVM的多类分类器对数据进行训练及分类,设计了一种能够应用于实际系统的汉字字符识别方法。The recognition on the Chinese character is always the difficult point in license plate recognition.A feasible methodology of the Chinese characters recognition is studied in this paper.In this paper,2D-Gabor and SVM are analyzed.After that,2D odd-symmetry Gabor wavelet is used as the filter function of characters,and the data is trained and classified by using multi-class SVM based on v-SVM.

关 键 词：二维奇对称Gabor小波 支持向量机 引入参数ν的支持向量机(ν-SVM) 多类SVM 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]