智能设计电表的自动识别系统  

Design of Automatic Recognition System of Intelligent Meter

作　　者：王晓员[1] 

机构地区：[1]合肥学院基础教学与实验中心,安徽合肥230022

出　　处：《广东技术师范学院学报》2013年第3期53-56,共4页Journal of Guangdong Polytechnic Normal University

摘　　要：针对电表表盘自动识别中,电表的图像歪斜、半个字符的识别、刻度对应的小数值推算,存在着运算精度与速度困难,建立了电表自动识别系统.其包括摄取电表图像的预处理与数字识别两个部分.对摄取的图像,先用纹理特征进行初步定位,然后用形态学对电表进行精确定位;进而采用Hough变换和小波分析,将图像倾斜校正,从复杂背景中去除干扰,分割出单一数字等一系列算法作为预处理.提出整个数字与非整个数字识别的不同算法,对前者采取小波分析,对后者采取基于像素与刻度比例的数字识别算法.最后对采集到的电表图像800幅,该系统得到单个数字识别率99.6%,整个电表图像识别率为98.1%.实验结果表明该系统具有较强的鲁棒性,具有很大的实用价值.In view of the automatic recognition of electric meter dial, meter image askew, half of the character recognition, scale corresponding small digital numerical difficult, to establish the automatic recognition system preprocessing and digital identification. For the image calculation, there are operation speed and accuracy is of meter. It consists of two parts： intake meter image on the uptake, first with the texture characteristics for preliminary positioning, then the meter for precise positioning by morphology; and then the Hough transform and wavelet analysis, the image tilt correction, from the complicated background to remove interference, segmentation of a single digital and/a series of algorithm as pretreatment. To put forward a different algorithm the whole digital and non whole digit recognition, for the former, using wavelet analysis, for the latter, take the digital recognition algorithm based on pixel and the proportion of the meter scale. For the collected meter image of 800, this system is a single digit recognition rate of 99.6%, the whole electric meter image recognition rate is 98.1%. The experimental results show that the system has strong robustness, it is of great practical value.

关 键 词：图像预处理 电表字段区域 单一数字分割 整个数字识别 局部数字识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]