Study On Monitoring Method Of Forest Fire Based On Geostationary Satellite

Fire is one of the main factors that cause instability in forest and grassland ecosystems.Forest and grassland fires have the characteristics of sudden bursts and rapid spread.Using geostationary satellite remote sensing technology can realize large-area synchronous and rapid monitoring of forest and grassland fires,also can provide data and decision support for fire fighting.This study was based on visible light multispectral and mid-wave infrared data(PMI)of Chinese GF-4 satellite,the Himawari-8 satellite data from Japan,with five experimental areas,which including Lijiang city of Yunnan province,Zhuhai city of Guangdong province,Pingshun county of Shanxi province of China,outer Baikal territory and Amur State of Russia.The study has been carried out on the fire points detection of Geostationary satellite data and change of fire points in time series data.The research content of this paper mainly included three aspects.On one hand,the algorithm of single-medium-wave infrared channel satellite data has been studied,the data of GF-4 PMI image and Himawari-8 satellite image in the medium-wave infrared channel were used,combining the brightness and reflectivity characteristics of typical objects in different research areas.On the other hand,the algorithm of fire detection of multi-thermal infrared channel has been studied.The multi-thermal infrared channel data of Himawari-8 satellite image were used to detect the fire points,and the accuracy was verified with MOD14 and MCD64A1 product.694 of Himawari-8 satellite images were detected by IDL program.Furthermore,the change of forest fire combustion state has been studied.The change trend of forest fire combustion state was extracted from 694 of Himawari-8 satellite images of fire point detection results and boundary results of burned area.The main research results and conclusions of this paper are as follows:(1)The adaptive threshold active fire detection algorithm for single-medium-wave infrared image of GF-4 PMI data,which proposed in this paper was tested and verified in five experimental areas.The overall detection accuracy of this algorithm was higher than 80.0%,and the comprehensive evaluation index of accuracy and missed detection rate was higher than 0.710.It shows that the algorithm is reliable.(2)The detection algorithms based on Himawari-8 satellite image using single-mediumwave infrared channel and multi-thermal infrared channel have been proposed separately in this paper.The detection and verification of the algorithm have been applied in Amur prefecture of Russia.The correct detection rate of single-medium-wave infrared channel was 78.9%,and the false detection rate was 23.4%.The correct detection rate of the multi-thermal infrared channel fire detection algorithm was 83.1%,and the false detection rate was 18.8%.Comparing with the two fire detection algorithms of Himawari-8 satellite images,the multi-thermal infrared channel fire detection algorithm has higher accuracy and more reliable results.(3)The fire point detection algorithm is used to obtain the fire point detection result of the continuous image by the high-frequency observation satellite image,which can reflect the combustion state of the fire field,such as the spread direction of the fire field and the duration of the fire point.Meanwhile,when a forest fire occurs,the spread trend and burning situation of forest fires can be reflected by boundary of burned area at certain intervals.Combining with the results of fire detection,forest fire can be simulated and predicted,which can provide data and decision support for forest fire fighting.