基于K均值聚类的变电站红外图像故障区域分割方法参数研究  

Research on parameters of fault region segmentation method for infrared images of substation based on K-means clustering

作　　者：肖懿 李伟绮 王韵 何渝霜 罗丹 XIAO Yi;LI Weiqi;WANG Yun;HE Yushuang;LUO Dan(State Grid Chongqing Electric Power Company Yongchuan Power Supply Branch,Chongqing 402160;State Grid Chongqing Electric Power Company Construction Branch,Chongqing 401100;State Grid Chongqing Electric Power Company Shinan Power Supply Branch,Chongqing 401336;School of Electrical&Information Engineering,Changsha University of Science and Technology,Changsha 410114)

机构地区：[1]国网重庆市电力公司永川供电分公司,重庆402160 [2]国网重庆市电力公司建设分公司,重庆401100 [3]国网重庆市电力公司市南供电分公司,重庆401336 [4]长沙理工大学电气与信息工程学院,长沙410114

出　　处：《电气技术》2024年第11期10-14,21,共6页Electrical Engineering

基　　金：湖南省自然科学基金(2024JJ6053)。

摘　　要：针对基于K均值聚类的变电站红外图像故障区域分割方法,本文首先简述K均值聚类算法应用于图像检测的原理;其次,通过K均值聚类进行红外图像故障区域提取,结果表明K均值聚类算法可以用于故障区域的分割识别;最后,分析K均值聚类算法参数对红外图像分割的影响。研究结果表明,红外图像的故障类别和初始分类点会对分割结果造成影响。在实际应用中,可以根据红外图像的特征进行初始设置,以提升图像分割的效率和准确性。This paper analyzed an fault region segmentation method for substation infrared images based on K-means clustering.Firstly,the principle of K-means clustering algorithm applied to image detection is introduced.Secondly,the fault region of infrared image is extracted by the K-means clustering algorithm,and the results show that the K-means clustering can be used for identification of fault region.Finally,the influence of K-means clustering algorithm parameters on infrared image segmentation is analyzed and discussed.The results show that the fault category and initial classification point of infrared image affect the accuracy of segmentation result.In practical application,the initial setting can be carried out according to the characteristics of infrared image to improve the efficiency and accuracy of image segmentation.

关 键 词：变电站 设备故障 红外图像 图像识别 K均值聚类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TM63[自动化与计算机技术—计算机科学与技术]