基于模板匹配的车牌识别及matlab实现 - 毕业论文（设计）

科类 理工类 编号（学号） 20082036

本科生毕业论文（设计）

基于模板匹配的车牌识别及matlab实现

License plate recognition based on template matching and matlab

implementation

云南农业大学 昆明 黑龙潭 650201

学 院： 基础与信息工程学院

专 业： 电子信息工程 年级： 2008级

论文（设计）提交日期： 2012年5月 答辩日期： 2012年5月

云南农业大学 2012年 05 月

目 录

摘 要 ......................................................................................................................................................... 1 ABSTRACT .................................................................................................................................................... 2 1 前 言 ......................................................................................................................................................... 3 2 车牌识别系统分析 ..................................................................................................................................... 4

2.1 车牌识别的目的 .............................................................................................................................. 4 2.2 车牌识别现状分析 .......................................................................................................................... 4 2.3 车牌识别的意义 .............................................................................................................................. 5 2.4 我国车牌分析 .................................................................................................................................. 6 3 车牌识别系统的原理及方法 ..................................................................................................................... 7

3.1车牌识别系统简述 ........................................................................................................................... 7 3.2 车牌图像处理 .................................................................................................................................. 8

3.2.1 图像灰度化 ........................................................................................................................... 8 3.2.2 图像二值化 ........................................................................................................................... 8 3.2.3 边缘检测 ............................................................................................................................... 9 3.2.4 图像闭运算 ......................................................................................................................... 10 3.2.5 图像滤波处理 ..................................................................................................................... 11 3.4 车牌字符处理 ................................................................................................................................ 12

3.4.1 阈值化分割原理 ................................................................................................................. 12 3.4.2 对车牌阈值化分割 ............................................................................................................. 14 3.4.3 字符归一化处理 ................................................................................................................. 14 3.5 字符识别 ........................................................................................................................................ 14

3.5.1 字符识别简述 ..................................................................................................................... 14 3.5.2 字符识别的分类 ................................................................................................................. 15 3.5.3 基于模板匹配的字符识别 ................................................................................................. 15

4 实验分析 ................................................................................................................................................... 16

4.1车牌定位过程及分析 ..................................................................................................................... 16 4.2 车牌字符识别 ................................................................................................................................ 19 4.3 车牌识别结果及分析 .................................................................................................................... 22 5 结 论 ....................................................................................................................................................... 24 参考文献 ....................................................................................................................................................... 25 致谢 ............................................................................................................................................................... 25 附录 ............................................................................................................................................................... 25

基于模板匹配的车牌识别及matlab实现

伏绍鸫

（云南农业大学基础与信息工程学院，昆明 650201）

摘 要

随着我国经济、交通的的快速发展，车牌定位系统以及车牌字符自动识别技术也逐渐受到人们的重视。车牌识别是对采集的车牌图像进行灰度变换、边缘检测、腐蚀以及平滑处理，最后在取得的大对象中移除小对象，由此提出了一种基于车牌纹理特征的车牌定位算法，最终实现对车牌的定位。车牌字符分割是为了以便后续对车牌模板进行匹配从而对车牌进行识别，考虑到我国车牌的结构构成所采用的字符并不是很多，由此本文采用了模板匹配的方法，对输出的图像和模板库里的模板进行匹配，通过处理后得到的图片与模板字符相减，得到的0越多那么就越匹配，然后对其个数进行累计并找出数值最大的，即为识别出来的结果。

经实验分析验证，本文所提出的整体方案有效可行，基于模板匹配的车牌识别技术在其识别正确率、速度方面具有独特的优势，前景广阔。

关键词：字符识别；模式识别；车牌定位；模板匹配；边缘检测

License plate recognition based on template matching and

matlab implementation

Fu Shaodong

(Yunnan Agricultural University information engineering, Kunming, 650201)

ABSTRACT

With the rapid development of China's economic and transportation, license plate positioning and automatic license plate character recognition system has been paid more attention gradually. The license plate recognition is a technology that uses the collection plate image to grayscale converting, edge detecting. Corrosion and smoothed, finally remove the small objects from the image, then put forward a algorithm of license plate recognition based on texture features. Finally, positioning the plate. License plate character segment is in order to facilitate the follow-up license plate template matching and license plate recognition. Considering the characters of our country’s license plate are not so many, Thus this thesis used the methods of template to matching the outputting image, the outputting image was subtracted from the template characters, the more 0 we get the more matches, then add up its numbers and find the largest, so it’s the identified .result.

Through the experlimental analysis and testing, the project this thesis put forward is effective and feasible, the license plate recognition technology based on template matching has the advantages on accuracy and rapidity, it has the vast potential for future development.

Key words: character recognition; pattern recognition; license plate location;