|
|
摘 要
图像识别在特定的环境下应用特别广泛。英文字母识别近年来发展迅速，已有大量的研究。英文字母识别关键点在特征提取方面，针对某一具体应用，所选择的特征往往直接影响到最终的识别率。
本文专门介绍了几种英文字母的识别及特征提取方法，重点介绍了如何提取模板匹配法来进行识别的方法，这个方法提取简单、区分度高、对用户不敏感，实现了有限的资源条件下的高速识别，同时保证了方法的良好用户适应性。
本论文研究的是基于模板匹配原则进行识别。论文分六个章节。第1章论述了选题意义与可行性分析以及字符识别的现状；第2章介绍了字符识别的几种方法；第3章主要介绍了本系统的主要算法；第4章介绍了本识别系统的实现，即本系统的预处理和识别方法；第5章介绍了程序运行结果及分析；第6章是结论。
关键词：图像识别；特征提取；模板匹配；
Abstract
The application of pictures is very widely in the particular environment. The recognition of English letters have a great development this years,and now there is a large number of research achievements. English letters in identifying key feature extraction, for a particular application, the feature selected directly affect the final often recognition.
This special introduces several English letters of the recognition and characteristic extraction method, this paper focuses on how to extract template matching method for identification method, this method is simple, high degree of distinction, user, realizing the limited resources under the condition of high recognition, the method of good adaptability to users.
In this paper, the study is based on the template matching principle character recognition. Thesis is divided into five chapters. The first chapter discusses the significance and feasibility analysis of topics as well as the character recognition of the status quo; The second chapter introduced several methods of character recognition; third chapter introduces the main algorithms of the system; the fourth chapter is devoted to the Recognition System, which is the current system and identify pre-treatment methods; The fifth chapter describes the results and analysis procedures; The sixth chapter is the conclusion.
Keywords：Picture identification; Feature extraction;Template matching
目 录
第1章 引言 1
1.1选题意义与可行性分析 1
1.2字符识别的现状 2
第2章 字符识别的几种方法 3
2.1模板匹配法 3
2.2人工神经网络法 4
2.3利用笔道关键点识别数字 5
2.4基于决策树的数字识别 6
2.5模糊模式识别 7
2.6基于灰度共生矩阵的文本图像识别 8
2.7用数学形态学处理字符 8
第3章 本系统的主要算法 10
3.1图像中的部分预处理 10
3.1.1二值化 10
3.1.2梯度锐化 10
3.1.3平滑去噪 11
3.1.4特征提取 11
3.2基于模板匹配原则的字符识别 12
3.2.1概述 12
3.2.2常用的模板匹配算法 14
3.2.3模板匹配的优缺点 17
3.2.4模板匹配设计思路 17
第4章 字符识别的实现 19
4.1本系统中的预处理 19
4.2本文中的模板匹配原则 20
4.3遇到的困难 22
第5章 程序运行结果及分析 23
第6章 结论 26
参考文献 27
致 谢 28
附 录 29
