摘要随着“一表一户”政策的普及以及计算机技术的发展，传统的人工抄表形式在效率以及准确率上都不再适合电力行业的大幅进步，于是计算机智能抄表技术应用而生。随着数字图像处理技术的发展，数字图像处理技术被运用到电能表自动抄表系统中，在电能表的图像识别技术中，图像分割起着举足轻重的作用。本论文在对非均匀光照条件下电能表图像的分割技术展开深入研究，利用数字图像处理技术并结合电能表图像特征解决非均匀光照对电能表图像分割的干扰，并实现电能表中有用信息的定位与分割问题。 论文的主要工作如下：
（1）对电能表图像进行预处理。首先对电能表进行灰度化处理后去除电表外边框。接着利用中值滤波平滑图像，减少噪声对图像的干扰。然后采用Sobel边沿检测提取G分量检测图像水平边缘，最后基于Hough变换结合区域特征进行倾斜校正。实现电能表图像的预处理。19544
（2）对电表图像进行几种常用的二值化方法并进行比较。采用数学期望法，全局阈值法，局部阈值法等对预处理过后的电能表图像进行二值化。对得到的二值化图像进行对比。
（3）采用一种改进的基于均值和方差的二值化方法对电表图像进行二值化处理。得出图像并与上述图像进行比较。用C++Builder软件进行处理。比较上述方法处理图像所需时间，得出结论。
关键词：图像预处理，二值化，数字图像处理
毕业论文设计说明书（论文）外文摘要
Title：Research under non-uniform illumination conditions meter image segmentation
Abstract
With the popularity of one family one electric meter policy and development of computer technology ,The traditional form of manual meter reading on the efficiency and accuracy are no longer suitable for substantial progress in the power sector,So smart metering technology computer generated .With the development of digital image processing technology , digital image processing technology is used to meter the automatic meter reading system,In the energy meter image recognition technology , image segmentation plays an important role .This paper seeks to non-uniform illumination conditions meter depth image segmentation technology research , using digital image processing technology combined with meter image features to solve the non-uniform illumination interference meter image segmentation and useful information to achieve energy meter positioning and segmentation .The main work is as follows :
（1）meter image preprocessing .First, the gray-scale processing power meter , then remove the meter outside the border.Then smooth the image using median filter to reduce noise on the image.Then extracted using Sobel edge detection level of the G component image edge detection ,Finally, based on Hough Transform area features tilt correction . Preprocessing meter image.
（2）the meter image of several commonly used binarization method and compared. Mathematical expectation method, global threshold, local threshold method for pre-processing after the meter image binarization . The binarized image obtained by comparison.
Based on the use of an improved method for the mean and variance of the binary image of the meter binarization . The image and the image obtained with the comparison. Treatment with C + + Builder software. Comparing the approach to the time required for image concluded .
Keywords: image preprocessing , binary , digital image processing
目录
Abstract - 3 -
1绪论 - 5 -
1.1课题背景及意义 - 5 -
1.2数字图像处理与识别技术的研究现状 - 6 -
1.3本文主要研究内容与各章节安排 - 7 -
2电表图像预处理 - 8 - 非均匀光照条件下电表图像分割技术研究:http://www.751com.cn/tongxin/lunwen_10947.html