融合图像识别与地理信息的电力设备送样监测技术  

Image recognition and geographic information integration of power equipment sample delivery monitoring technology

作　　者：刘志伟[1] 宁克[1] 刘星廷 侯滨[1] 王海旗 LIU Zhiwei;NING Ke;LIU Xingting;HOU Bin;WANG Haiqi(Materials Branch,State Grid Shanxi Electric Power Co.,Ltd.,Taiyuan 030021,China;Electric Power Science Research Institute,State Grid Shanxi Electric Power Co.,Ltd.,Taiyuan 030021,China)

机构地区：[1]国网山西省电力公司物资分公司,山西太原030021 [2]国网山西省电力公司电力科学研究院,山西太原030021

出　　处：《电子设计工程》2025年第3期106-110,共5页Electronic Design Engineering

基　　金：国网山西省电力公司科技项目(B9QD-300009601-00001)。

摘　　要：为提高送样过程的透明度和安全性,防止电力设备遭到非法干扰与损伤,优化送样过程的准确性及质量,文中提出了一种基于图像识别和地理信息的电力设备采样监测技术。该技术采用基于深度学习算法的图像特征提取技术对视频监控进行目标识别,并通过SAR快速获取目标特征及其相对位置,实现对抽检物体的识别和全流程监控。同时,采用椭圆曲线技术构建基于地理位置信息的视频加密策略,实现了对监控视频的安全保护。以电力设备抽检视频为数据样本进行的验证分析结果表明,所提检测方法的相关指标比同类型方法提升约10%,识别结果均保持在85%以上,可以准确地对抽检全过程进行监测,并能够分析处理抽检过程中的位置变化信息,具有更高的安全性。To improve the transparency and safety of the sample sending process,prevent illegal interference and damage to power equipment,and optimize the accuracy and quality of the sample sending process,this paper proposes a power equipment sampling monitoring technology based on image recognition and geographic information.This technology uses image feature extraction technology based on deep learning algorithms for target recognition in video surveillance,and quickly obtains target features and their relative positions through SAR,achieving recognition of sampled objects and full process monitoring.At the same time,the use of elliptic curve technology to construct a video encryption strategy based on geographic location information has achieved security protection for surveillance videos.The validation analysis results using power equipment sampling videos as data samples show that the relevant indicators of the proposed detection method are improved by about 10%compared to similar methods,both of which remain above 85%.It can accurately monitor the entire sampling process and analyze and process position change information during the sampling process,with higher security.

关 键 词：电力设备送样 深度学习 图像识别 地理信息 数据加密 椭圆曲线 

分 类 号：TP391.44[自动化与计算机技术—计算机应用技术] TN929.5[自动化与计算机技术—计算机科学与技术]