基于SVM多类分类器的字符识别  

Recognition of vehicle license-plate based on support vector machine

作　　者：李雪花[1] 许姜涤宇 于安军[1] 杜宇人[1] 

机构地区：[1]扬州大学信息工程学院,江苏扬州225127

出　　处：《信息技术》2016年第1期20-22,26,共4页Information Technology

基　　金：国家自然科学基金资助项目(51273172);江苏省创新创业训练计划立项项目(201511117054Y)

摘　　要：针对字符识别相似字符识别能力差,速度慢等问题,提出基于支持向量机的识别方法。首先提取预处理后的字符特征,然后进行降维处理,再设计多类分类器并进行训练,最后选取合适的核函数进行识别。实验表明,该方法识别率高达97.65%,且具有适应性强和效率高的特点,具有较高的实用价值。In view of the character recognition of similar characters recognition ability is poor, the speed is slow, the recognition method based on support vector machine （SVM） is put forward. After preprocessing of character feature extracting first, then it reduces the dimension, and then designs muhi- class classifier and training. Lastly, it selects the appropriate kernel functions for identification. The experiments show that the method recognition rate is as high as 97.65%. And it has the characteristics of strong adaptability and high efficiency, has high practical value.

关 键 词：字符识别 支持向量机 多类分类器 核函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]