基于OpenCV的变电站仪表识别方法研究  

Research on the Recognition Method of Substation Instruments Based on OpenCV

作　　者：杨典 李小燕 刘培焱 刘丰硕 YANG Dian;LI Xiao-yan;LIU Pei-yan;LIU Feng-shuo(Electrical and Electronic Teaching Experiment Center,Xi’an Jiaotong University,Xi’an 710049,China;Yunfu Power Supply Bureau,Guangdong Power Grid Co.,Ltd.,Yunfu 527300,China)

机构地区：[1]西安交通大学电工电子教学实验中心,西安710049 [2]广东电网有限责任公司云浮供电局,云浮527300

出　　处：《自动化与仪表》2022年第4期75-80,共6页Automation & Instrumentation

基　　金：2021年国家级大学生创新训练项目(202110698249)。

摘　　要：机器视觉和图像处理的发展为工业仪表识别的自动化智能化提供了基础。该文基于OpenCV的图像处理功能分别对两种不同种类的仪表进行示数识别。对于数字式仪表,利用图像处理分割表盘单个数字图片后,运用卷积神经网络进行数字识别;对于指针式仪表,先进行透视变换,获得正规圆形表盘后通过直线拟合以确定指针旋转角度,进而计算仪表示数。实际测试表明,采用该文的方法对实际拍摄仪表照片进行识别,准确度高,速度快,在变电站等工业仪表识别方面具有一定的应用价值。The development of machine vision and image processing provides the basis for the automation and intelligence of industrial instrument recognition. Based on the image processing function of OpenCV,this paper recognizes the readings of two different kinds of instruments. For the digital instrument,the single number picture of the dial is segmented by image processing,and the convolution neural network is used for number recognition;For the pointer instrument,the perspective transformation is carried out first to obtain the normal circular dial,then the rotation angle of the pointer is determined through straight-line fitting,and finally the reading of the instrument is calculated. The actual test shows that using this method to identify the actual instrument photos has high accuracy and high speed.It has a certain application value in the identification of industrial instruments such as substations.

关 键 词：OPENCV 图像识别 数字式仪表 指针式仪表 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TH70[自动化与计算机技术—计算机科学与技术]