Automatic infrared image recognition method for substation equipment based on a deep self-attention network and multi-factor similarity calculation

作     者：Yaocheng Li Yongpeng Xu Mingkai Xu Siyuan Wang Zhicheng Xie Zhe Li Xiuchen Jiang Yaocheng Li;Yongpeng Xu;Mingkai Xu;Siyuan Wang;Zhicheng Xie;Zhe Li;Xiuchen Jiang

作者机构：School of Electronic Information and Electrical EngineeringShanghai Jiao Tong University200240 ShanghaiP.R.China State Grid Shandong Electric Power Company Jinan Power Supply Company250001 ShandongP.R.China Maintenance&Test Center of EHV power Transmission CompanyChina Southern Power Grid510000 GuangdongP.R.China 

出 版 物：《Global Energy Interconnection》 (全球能源互联网（英文版）)

年 卷 期：2022年第5卷第4期

页      面：397-408页

核心收录：

学科分类：080802[工学-电力系统及其自动化] 0808[工学-电气工程] 080901[工学-物理电子学] 0809[工学-电子科学与技术（可授工学、理学学位）] 08[工学] 080203[工学-机械设计及理论] 080401[工学-精密仪器及机械] 0804[工学-仪器科学与技术] 0802[工学-机械工程] 0803[工学-光学工程] 

基　　金：This work was supported by National Key R&D Program of China(2019YFE0102900) 

主　　题：Substation equipment Infrared image intelligent recognition Deep self-attention network Multi-factor similarity calculation 

摘      要：Infrared image recognition plays an important role in the inspection of power *** technologies dedicated to this purpose often require manually selected features,which are not transferable and interpretable,and have limited training *** address these limitations,this paper proposes an automatic infrared image recognition framework,which includes an object recognition module based on a deep self-attention network and a temperature distribution identification module based on a multi-factor similarity ***,the features of an input image are extracted and embedded using a multi-head attention encoding-decoding ***,the embedded features are used to predict the equipment component category and *** the located area,preliminary segmentation is ***,similar areas are gradually merged,and the temperature distribution of the equipment is obtained to identify a *** experiments indicate that the proposed method demonstrates significantly improved accuracy compared with other related methods and,hence,provides a good reference for the automation of power equipment inspection.