基于多源遥感数据的电化学储能电站火灾监测方法  

Fire monitoring method for electrochemical energy storage power plants based on multi-source remote sensing data

作　　者：闫观捷 YAN Guanjie(The Department of Basic Education,Shanghai Urban Construction Vocational College,ShangHai 201499,China)

机构地区：[1]上海城建职业学院基础教学部,上海201499

出　　处：《自动化与仪器仪表》2025年第2期98-101,共4页Automation & Instrumentation

基　　金：上海城建职业学院2024年度校级科研项目资助(AA-02-2024-01-16-47);上海城建职业学院职业本科强师计划项目(AA-05-2024-02-17-01)。

摘　　要：电化学储能电站火灾存在监测难度大、响应速度慢以及潜在危害严重等问题,这些问题容易引发火灾初期的难以察觉、火势迅速蔓延以及人员伤亡和财产损失等严重后果,设计基于多源遥感数据的电化学储能电站火灾监测方法。选取两个时效性高的遥感监测数据Himawari-8、FY-4A与两个稳定性好的遥感监测数据MODIS、NOAA-19作为数据来源,为提升遥感原始影像的分析精度,实施云检测,通过同一地区、不同时间但无云覆盖的遥感图像数据替换云层区域数据。应用时空自适应反射率融合模型实施四种遥感监测数据的时空融合,生成同时具有高稳定性与高时效性的遥感影像。利用绝对火点像元判定法实施融合后遥感影像的火点像元判定,实现电化学储能电站火灾监测。检测并排除太阳光反射导致的虚假火点。测试结果表明,设计方法的监测结果与实际情况十分贴近,监测较为准确,在长期运行下(6~18个月),设计方法的遗漏次数始终为零。Electrochemical energy storage power plant fires have problems such as difficult monitoring,slow response speed,and serious potential hazards.These problems can easily lead to serious consequences such as difficulty in detecting fires in the early sta-ges,rapid spread of fire,and casualties and property losses.Therefore,a fire monitoring method for electrochemical energy storage power plants based on multi-source remote sensing data is designed.Two highly timely remote sensing monitoring data Himawari-8 and FY-4A,as well as two stable remote sensing monitoring data MODIS and NOAA-19,were selected as data sources to improve the analysis accuracy of remote sensing raw images.Cloud detection was implemented by replacing cloud area data with remote sens-ing image data from the same region but with no cloud coverage at different times.Applying a spatiotemporal adaptive reflectance fu-sion model to implement spatiotemporal fusion of four types of remote sensing monitoring data,generating remote sensing images with both high stability and high timeliness.Using absolute fire point pixel determination method to implement fire point pixel determina-tion of fused remote sensing images,achieving fire monitoring of electrochemical energy storage power plants.Detect and eliminate false fire points caused by sunlight reflection.The test results show that the monitoring results of the design method are very close to the actual situation,and the monitoring is relatively accurate.Under long-term operation(6 months to 18 months),the number of o-missions in the design method is always zero.

关 键 词：多源遥感数据 时空融合 电化学储能电站 火灾监测 虚假火点排除 

分 类 号：P237[天文地球—摄影测量与遥感] TP274[天文地球—测绘科学与技术]