基于图像识别的变电站虚端子连接校核技术研究  

Research on Virtual Terminal Connection Verification Technology for Substation Based on Image Recognition

作　　者：焦晓晨 JIAO Xiaochen(State Grid Lanzhou Power Supply Company,Lanzhou,Gansu 730000,China)

机构地区：[1]国网兰州供电公司,甘肃兰州730000

出　　处：《自动化应用》2024年第22期91-93,96,共4页Automation Application

摘　　要：变电站配置是极为复杂的系统,要想实现对变电站的检验,需要引入计算机技术,以解决装置核验准确率低、人工校验不全等问题。鉴于变电站校验的新要求,引入了图像识别方法,应用R-CNN回归算法、SCD虚连接算法以及轮廓追随算法,构建校核变电站虚端子连接工具,通过识别文件、虚连接对以及分割训练,完成对变电站虚端子的校核,识别变电站中的图像和文件,以期从变电站中筛选出不适配信息,出具变电站校核报告。Substation configuration is an extremely complex system.In order to achieve inspection of substations,computer technology needs to be introduced to solve problems such as low accuracy of device verification and incomplete manual verification.In view of the new requirements for substation verification,this paper introduces an image recognition method,applies R-CNN regression algorithm,SCD virtual connection algorithm,and contour following algorithm to construct a virtual terminal connection tool for verification substations.By identifying files,virtual connection pairs,and segmentation training,the verification of virtual terminals in substations is completed.Images and files in substations are identified,in order to filter out incompatible information from substations and issue a substation verification report.

关 键 词：图像识别 变电站虚端子 连接校核 

分 类 号：TM63[电气工程—电力系统及自动化] TP391.41[自动化与计算机技术—计算机应用技术]