摘要： 随着社会的进步与发展，快递将越来越多，为了更好地管理快递，智能快递系统将会是是发展的方向，而邮编自动识别系统是智能快递系统之中一个重要的环节，它主要包含2个重要技术：邮编字符分割和邮编字符识别。本文针对邮编自动识别系统中关键技术之一的邮编字符分割部分做了深入的研究。29423
首先，要想将邮编图像正确地分割，必须得到质量较好的邮编二值图像，因此本文对邮编字符分割前的预处理进行了较深入研究，尤其是邮编图像二值化后的去噪处理，包括开闭运算、去除边缘、去除圆点等。
其次，在字符分割算法中，给出两种分割算法，一种是固定边界法，另一种是垂直投影法，并对邮编字符进行了分割。实验结果表明，这两种方法可以对邮编字符进行分割，但效果都不是很理想。
然后，在分析了这两种方法的优缺点后，将这两种方法相结合得到基于灰度直方图峰谷法的分割，实验表明，该方法可以得到更好的字符分割效果，尤其对在除噪后仍然有少量噪声的邮编也可达到较好的效果。
毕业论文关键词： 字符分割；二值化；预处理；投影法
The Development of the System of Express Automatic Recognition—Characters Checking and Cutting
Abstract： There will be more and more express with the development of society. Intelligent Express System will be the direction of development. System of Express Automatic Recognition is an important part of the Intelligent Express System. It is contain of 2 key technologies： the pision of postcodes and the recognition of postcodes. This paper studied at character recognition technology which is one key technology of the System of Express Automatic Recognition.
At first， the better quality postcodes is needed for properly pide the postcodes. So this paper thoroughly studied the pretreatment before the pision， especially the wipe of the noise after the binarization including open arithmetic， close arithmetic， wipe of the edge， and wipe of the round spot etc.
Secondly， there are two methods to pide the postcodes. The first is the method of invariable boundary and the other is the method of vertical shadow. At the same time， make the pision in the two methods separately. The tests indicate that both of the two methods can make the pision， but both of them can not do it well.
Finally， combine the two methods and reach the improving method of vertical shadow after analyze merit and demerit of the two methods， The tests indicate that this method make a better pision especially to the postcodes which still have some noise after wiping the noise.
Keywords： the pision of postcodes；binarization；pretreatment；the method of shadow
目录
摘要 i
Abstract i
目录 v
1 绪论 1
1.1 邮编识别技术的应用背景及前景 1
1.1.1 21 世纪邮编识别的发展 1
1.1.2 邮编识别的领域 2
1.2 邮编识别系统的工作原理及组成 2
1.3 字符分割的发展现状 3
1.4 本文研究主要内容 4
1.5 本文的结构 4
2 图像处理简介 6
2.1 图像采集 6
2.2 二值图像 7
2.3 灰度图像 7
2.4 灰度直方图 8
2.5 文字切分 9 Matlab邮编自动识别系统的开发数字检测与切分研究:http://www.751com.cn/jisuanji/lunwen_24664.html