基于卫星影像的输电线路周边地表隐患识别方法  

Satellite image-based method for extracting potential surface dangers along power transmission lines

作　　者：邹绍悦 刘朔 夏昊 ZOU Shaoyue;LIU Shuo;XIA Hao(University of Chinese Academy of Sciences,Beijing 100049,China;Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China)

机构地区：[1]中国科学院大学,北京100049 [2]中国科学院空天信息创新研究院,北京100094

出　　处：《中国科学院大学学报（中英文）》2023年第6期778-787,共10页Journal of University of Chinese Academy of Sciences

基　　金：国家重点研发计划(2020YFF0400405)资助。

摘　　要：我国输电线路面临着多种周边地物影响带来的安全隐患,线路隐患排查对确保国家电力可靠供应及电力安全具有重要意义。当前各种线路隐患排查方法面临不同的困难,而卫星遥感技术可提供低成本、非接触的快速解决方案。探索使用卫星遥感影像实现输电线路周边地表隐患识别与提取的解决方案,提出一种针对输电线路周边地表隐患识别提取的方法,使用多类别残差修正模块进行结果修正,减少小尺度目标识别不佳及类别边缘误判的情况,使输电线路周边地表隐患识别效果得到提升。Power transmission lines in China are faced with safety hazards brought about by a variety of surrounding features,and line hazard investigation is of great significance to ensure reliable national power supply and power security.At present,various methods of line hidden danger investigation face different difficulties,and satellite remote sensing technology can provide a low-cost,non-contact and rapid solution.In this paper,we use satellite remote sensing technology to realize the identification and extraction of hidden dangers on the ground along transmission lines.A method is also proposed for the identification and extraction of hidden hazards along transmission lines,using multi-category residual refinement module for result correction,which reduces the cases of poor recognition of small-scale targets and category edge misclassification,and lays the methodological foundation for the practical application of satellite remote sensing technology to the investigation of hidden dangers along power transmission lines.

关 键 词：输电线路安全 隐患识别 U-net RRM_Mc 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]