基于图像的电能表数字识别算法研究
毕业论文关键词 图像处理和识别 图像的二值化 图像的细化 模板匹配法
Title Comparative Study of Several Kinds of Commonly Used Recognition Algorithm of Handwritten Numbers
Abstract Watt-hour meter is a device which will be used in every household. In most cases we use manual reading to bill electricity consumption,this method has a low automation degree, data acquisition waste a lot of human resources, and efficiency is very low. So it is necessary to design an effective method to improve the efficiency of reading the data of watt-hour meter. Image processing and recognition is a new technology that developed in the last 20 years.This paper mainly introduces the method of using the template matching method to recognize the watt-hour meter. In order to achieve the digital identification of watt-hour meter, the work of this paper is pided into three parts: image preprocessing, feature extraction of digital image and image recognition. Image preprossing is image gray processing,image binarization and fringe patten sharpening mainly;feature extraction of digital image recognition is mainly using digital contour structure characteristic and the statistical characteristics to select feature vector, then use template matching method for digital identification and to be accurately identify the digital image. After many times of recognition, the method used in this paper can achieve the function of the digital identification of the watt-hour meter, which has better recognition effect.
Keywords image processing and Recognition Binary image Image Refinement template matching method
目次
1引言.1
1.1研究背景及意义1
1.2仪表识别的发展历史1
1.4论文主要工作和内容安排.2
1.5开发平台简介2
2数字图像基础.2
2.1数字图像处理概述4
2.2图像文件4
2.3BMP数字图像4
2.4位图及文件结构分析5
2.5本章小结7
3图像的预处理.8
3.1图像的灰度处理8
3.2图像的二值化.10
3.3图像的细化.11
3.4本章小结.12
4特征识别13
4.1数字仪表识别概述13
4.2特征提取13
4.3本章小结.15
5数字识别算法16
5.1模板匹配算法16
5.2BP神经网络算法17
5.3实验结果和分析.19
6总结.21
结论.22
致谢.23
参考文献24
1 引言
1.1 研究背景及意义 社会不断进步，科技不断地发展，在当今我们的生活^吹冰!文`论^文'网www.chuibin.com、工作中，处处可见高科技的身影。人们不再像从前那样被重复而又繁琐的机械式操作所牵扯，而是采用机器或者计算机来完成这些工作。在众多的机器和计算机技术中，机器视觉识别技术是受到广泛应用的；它可以实现计算机识别以及处理图像，还可以分析比较，替人们解决了许多繁琐的工作。 光学字符识别（OCR）最早可以追溯到 20 世纪 20 年代，现在已发展成为模式识别研究的一个重要分支。光学字符识别通过扫描、摄像等光电非接触方式，将字符信息转化成图像，应用字符识别技术快速、准确的对图像进行识别，与传统方法相比，极大地减少了资源浪费，提高了工作效率。 当前大多数仪表都是电子式，比如本文所识别的电能表，仪表中的数字是由二极管或者液晶屏显示出来的，属于印刷字体，非常容易进行识别。所以我们发展图像处理技术，实现自动化读取仪表中的信息是工业自动化中的重要里程碑，对工业发展具有重要价值。