字符识别应用与开发研讨.docx

字符识别应用开发 专业：电子信息工程 摘要 模式识别在现今已经成为全世界热门的研究课题，也被越来越广泛的应用。而字符识别系统作为模式识别的一个重要分支，也受到了很高的重视度。计算机中的模式识别，其实就是要让计算机学会“阅读”，提取里面的信息。而在模式识别中，字符识别则是其中应用最广泛的一个。在字符识别算法、系统不断完善的同时，也有一些不足。我们在这里主要介绍了字符识别的背景，研究字符识别的基本方法，探讨了字符识别的过程，研究图像处理及识别过程的算法，采用了基于调节因子的灰度均值二值化算法对图像进行二值化，将数字图像进行准确分离为数字前景和背景。本文还采用了投影法对字符串进行分割，很好的解决了字符分割的问题。分割结束后将字符归一化为标准大小，然后提取字符的多组特征向量，并通过模板匹配法完成了系统的实现。还根据具体的票据字符识别情况提出一些注意事项及问题。 关键词：字符识别，ocr，模版匹配，神经网络,matlab Application development of character recognition Major: Electronic information engineering Abstract Pattern recognition in nowadays has become a hot research topic from all over the world,and it has been widely used. The character recognition system ,as an important branch of pattern recognition, also brought to high attention. In the computer pattern recognition, in fact ,is to make computers learn to read, and extract the information inside. In pattern recognition, character recognition is one of the most patern which is widely used. With the character recognition algorithm, system constantly have improved , there are also some shortcomings. Here we mainly introduced the background of the character recognition, the research and the basic methods of character recognition, and discusses the character recognition process, study the process of image processing and recognition algorithm, and adopted the grayscale average binarization algorithm based on regulating factor for image binarization, the digital image accurately separated into foreground and background. This paper also adopted the projection of string segmentation, good solve the problem that the character segmentation. After the split it will becomes standard big characters。` Key words:character recognition,ocr, tempte match,neural networks,matlab 目录 第一章前言 1 第二章字符识别概述 2 2.1研究背景及意义 2 2.2票据字符识别研究现状 2 2.3字符识别方法 3 2.2.1模板匹配法 4 2.2.2 BP神经网络字符识别 4 2.4 影响因素，重要指标 6 第三章字符识别系统设计 7 3.1系统的组成 7 3.2系统的模块 7 3.3图像预处理模块设计 7 3.4数字识别模块的设计 11 3.5界面设计 11 第四章数字字符识别系统的实现 12 4.1图像数据读取 12 4.2图像的二值化 12 4.3去除离散杂点噪声 13 4.4字符