期刊文献+

基于激光非线性成像技术的配网数字化作业安全远程红外监测系统

Research on Remote Infrared Monitoring System for Digital Operation Safety of Distribution Network Based on Laser Nonlinear Imaging Technology
原文传递
导出
摘要 为准确识别配网数字化作业人员,提高配网数字化作业安全远程红外监测效果,研究基于激光非线性成像技术的配网数字化作业安全远程红外监测系统。系统的硬件设计为感知层、网络层、应用层和用户层4层,利用感知层的SF6激光非线性摄像头采集配网数字化作业全过程红外图像,使用网络层的Internet将采集的图像传输到应用层;根据SF6激光非线性摄像头结构和成像原理,设计配网数字化作业安全远程红外监测终端硬件结构;系统的软件部分,设计配网数字化作业安全远程红外监测流程;基于卷积神经网络算法搭建多尺度金字塔,提取作业人员特征图谱,通过快速局部拉普拉斯滤波变换增强图像的像素特征;基于Softmax函数识别配网数字化作业人员的危险区域侵入行为,实现配网数字化作业安全远程红外监测。试验结果表明:该系统的图像增强处理能力和作业人员识别效果较理想,并能及时、有效地预警作业人员的危险区侵入行为,预警用时始终低于0.06 s,且网络传输综合性能较为优异,数据丢包量最高为5000个,提高了配网数字化作业安全远程红外监测效果。 In order to accurately identify the digital operators of distribution network and improve the effect of remote infrared monitoring of digital operation safety of distribution network,a remote infrared monitoring system of digital operation safety of distribution network based on laser nonlinear imaging technology is studied.The system entails the design of a comprehensive digital image acquisition process covering the SF6 layer and the infrared application layer.Furthermore,it encompasses the hardware configuration of a remote infrared monitoring terminal tailored for the digital operation safety of distribution networks,aligned with the structural and imaging principles of SF6 laser nonlinear cameras.The software component orchestrates the remote infrared monitoring process,incorporating a convolutional neural network algorithm to construct a multi-scale pyramid for operator feature extraction.Pixel features are then enhanced through fast local Laplace filter transformation.Intrusion behavior detection within hazardous areas is achieved via softmax function,facilitating safe remote infrared monitoring of distribution network digital operations.The experimental results?show that the image enhancement processing ability and operator recognition effect of the system are ideal,and can timely and effectively warn the intrusion behavior of operators in the dangerous area.The warning time is always less than 0.06s,and the comprehensive performance of network transmission is excellent.The maximum amount of data packet loss is 5000,which improves the effect of safe remote infrared monitoring of digital operation in distribution network.
作者 高军伟 安志国 俞思帆 李英锋 李颂杰 洪海生 Gao Junwei;An Zhiguo;Yu Sifan;Li Yingfeng;Li Songjie;Hong Haisheng(Guangzhou Nansha Power Supply Bureau of Guangdong Power Grid Co.,Ltd,Guangzhou511470,Guangdong,China;Guangzhou Power Supply Bureau of Guangdong Power Grid Co.,Ltd.,Guangzhou510620,Guangdong,China)
出处 《应用激光》 CSCD 北大核心 2024年第3期171-178,共8页 Applied Laser
基金 中国南方电网科技项目“基于5G的配网数字化作业工作量智慧监管系统”(086400KK52200003)(GZHKJXM20200049)。
关键词 激光非线性 成像技术 配网数字化 作业安全 远程红外监测 卷积神经网络 laser nonlinearity imaging technology distribution network digitization operation safety remote infrared monitoring convolutional neural network
  • 相关文献

参考文献14

二级参考文献137

共引文献188

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部