| 字符特征的可分性量化分析 |
| |
| 引用本文： | 程卫东,王天杨.字符特征的可分性量化分析[J].北方交通大学学报,2010(1):140-144. |
| |
| 作者姓名： | 程卫东 王天杨 |
| |
| 作者单位： | [1]北京交通大学机械与电子控制工程学院,北京100044 [2]辽宁工程技术大学大型工矿装备实验研究中心,辽宁阜新123000 |
| |
| 基金项目： | 辽宁工程技术大学大型工矿装备实验研究中心基金项目资助（M0710020）;北京交通大学科技基金项目资助（2006XM028） |
| |
| 摘 要： | 选择字符识别的特征时要进行特征的可分性量化分析．文中分析了5种在汽车号牌识别中用到的字符特征提取方法，进行了可分性测度量化计算．采用主成分分析法对特征降维处理，可以简化可分性测度的计算，使结果更加直观清晰．测度图谱的表示方法，能较好地对字符特征进行确定及取舍．采用其中3种特征的组合，即有较好的可识别性，又可降低计算工作量．
|
| 关 键 词： | 模式识别 可分性测度 主成分分析 测度图谱 特征提取 |
Quantitative Analysis of Recognizability of Character Features |
| |
| CHENG Weidong,WANG Tianyang.Quantitative Analysis of Recognizability of Character Features[J].Journal of Northern Jiaotong University,2010(1):140-144. |
| |
| Authors: | CHENG Weidong WANG Tianyang |
| |
| Institution: | 1. School of Mechanical and Electronic Control Engineering, Beijing Jiaotong University, Beijing 100044, China; 2. Open Research Fund Program of the Experiment and Study Center of Heavy Engineering and Mining Equipments, Liaoning Technical University, Fuxin Liaoning 123000, China) |
| |
| Abstract: | Recognizability of character Features must be analyzed quantificationally when selecting fea- ture for characters recognition. In this paper, five types extracting method of characters features on vehicle license plate are analyzed in detail, and recognizability measurements are calculated quantificationally. Primary component analysis （PCA） method is used to reduce dimension of features, and calculation of recognizability measurements are predigested so that result figure becomes much more direct and perspicuous. The use of measurement spectrum is a good way to validate and accept or reject fea- tures of characters for the fast classification and recognition of characters on vehicle license plate. Using combined optimum features may improve accurate recognition ratio and reduce calculation workload. |
| |
| Keywords: | pattern recognition recognizability measurement PCA measurement spectrum extraction of features |
| 本文献已被 维普 等数据库收录！ |
