电能表表码识别图像预处理研究
Research of Image Preprocessing for Digital Record Recognition of Power Meter
陈佳莹1
作者信息
- 1. 广东电网有限责任公司汕头供电局 广东 汕头 515000
- 折叠
摘要
电能表批量轮换工作涉及大量表计,这要求在装拆现场能够准确、快速采集新旧电能表编号和旧电能表对应的表码信息,以便后期系统档案的维护.将OCR技术应用到电能表轮换工作中,可以在不添加特殊终端的情况下快速采集到表码信息.进行OCR字符识别前对图像进行必要的预处理能大大提高字符识别的准确率.本文针对电能表表码采集这一场景,介绍图像预处理中二值化、去除噪点和表码主体提取3个步骤的具体细节.
Abstract
The project of replacing power meters involves a large batch of power meters,which requires to markdown the identification number of the new and old meters and the digital record information of the related old meters at assembly and disassembly site accurately and quickly,for subsequent maintenance of the archives on the system.Applying OCR technique in the project of replacing power meters can quickly collect the digital record information without using any special devices.The necessary preprocessing of the images before OCR character recognition will greatly increase the accuracy of character recognition.This paper introduces the details of binarization,noise removal and digit extraction in image preprocessing based on the scene of digital record collection of power meter.
关键词
表码采集/图像预处理/二值化/去除噪点/表码主体提取Key words
digital record collection/image preprocessing/binarization/noise removal/digit extraction引用本文复制引用
出版年
2024
NETL
NSTL
万方数据