基于光谱显著性的复合绝缘子识别算法  

Composite Insulator Identification Algorithm Based on Spectrum Saliency

作　　者：黄力 廖梓榕[2] HUANG Li;LIAO Zirong(Hubei Provincial Engineering Technology Research Center for Power Transmission Line,Hubei Yichang 443002,China;College of Electrical Engineering&New Energy,China Three Gorges University,Hubei Yichang 443002,China)

机构地区：[1]湖北省输电线路工程技术研究中心,湖北宜昌443002 [2]三峡大学电气与新能源学院,湖北宜昌443002

出　　处：《电工材料》2023年第3期1-4,共4页Electrical Engineering Materials

基　　金：湖北省输电线路工程技术研究中心开放课题(2019KXL08)。

摘　　要：针对背景纹理复杂、相近色干扰以及不同光照环境下复合绝缘子不能准确分割的情况,提出一种基于光谱显著性的复合绝缘子识别算法。将该方法分别与H通道法、色差法、频率调谐法及色相明度融合法比较,五者平均分割准确率分别为57.55%、48.47%、14.80%、73.05%和82.73%,分别提高了25.18%、34.26%、67.93%和9.68%,表明该方法拥有较高的准确性和适应性,为后续绝缘子特征提取与故障分析提供了有利参考。In view of the complex background texture,similar color interference and the different lighting environments,composite insulators cannot be accurately segmented.The composite insulator recognition algorithm based on spectrum saliency is proposed.By comparing the method with the H-channel method,the color difference method,the frequency-tuned method and the hue brightness fusion method respectively,the average segmentation accuracy is 57.55%,48.47%,14.80%,73.05%and 82.73%,respectively,the average segmentation accuracy is reduced by 25.18%,34.26%,67.93%and 9.68%.The results show that the method has high accuracy and adaptability,which provides a favorable reference for the feature extraction and fault analysis of composite insulators.

关 键 词：复合绝缘子 背景复杂 光照不均 光谱显著性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TM216[自动化与计算机技术—计算机科学与技术]