基于图像识别的电力作业现场安全监督方法  

Research on Safety Supervision Method for Electric Power Operation Site Based on Image Recognition

作　　者：杨张慧 YANG Zhanghui(Guodian NARI Technology Co.,Ltd.,Nanjing 210000,China)

机构地区：[1]国电南瑞科技股份有限公司,江苏南京210000

出　　处：《电工技术》2025年第3期146-149,共4页Electric Engineering

摘　　要：为解决传统电力作业现场中存在的误报和漏报现象较多、图像处理效果较差等问题,提出一种基于图像识别技术的电力作业现场安全监督方法,以确保电力现场作业安全、排除安全隐患。首先,构建电力现场图像采集平台,对电力作业现场设备进行图像采集,并对采集的图像进行去噪处理,以提高图像的清晰度和准确性。然后,采用方向梯度直方图技术对电力作业现场图像进行特征提取,从而实现对电力作业现场的分类和识别。最后,建立电力作业现场安全监督系统,将采集的电力作业现场图像特征输入其中,并结合异物入侵检测算法对电力作业现场进行实时的安全监控和报警,从而提高电力作业现场的安全性。实验表明,这种方法能够有效地处理电力作业现场图像,对威胁电力作业现场安全的因素进行检测和报警,更好地实现电力作业现场的安全监督。To address the issues of frequent false positives and omissions in traditional power operation sites,as well as poor image processing performance,a safety supervision method for power operation sites based on image recognition technology is studied to ensure the safety of power operation sites and eliminate safety hazards.Firstly,build a power on-site image acquisition platform,collect images from equipment on the power operation site,and denoise the collected images to improve their clarity and accuracy.Then,the directional gradient histogram technology is used to extract features from the image of the power operation site,thereby achieving classification and recognition of the power operation site.Finally,establish a safety supervision system for the power operation site,input the collected image features of the power operation site into it,and combine foreign object intrusion detection algorithms to conduct real-time safety monitoring and alarm on the power operation site,thereby improving the safety of the power operation site.Experiments have shown that this method can effectively process images of power operation sites,accurately detect and alarm threats to the safety of power operation sites,and better achieve safety supervision of power operation sites.

关 键 词：图像识别 图像去噪 电力作业现场 安全监督 HOG特征提取 

分 类 号：TN249[电子电信—物理电子学]