Research On Fire Point Monitoring With High Time And High Spatial Resolution Based On Multi-source Remote Sensing Data

In recent years,the frequent occurrence of forest fires around the world not only poses a huge threat to the earth 's ecosystem,but also has a serious impact on human life.Due to the advantages of high update frequency,wide monitoring range and high work efficiency,satellite remote sensing monitoring of forest fire spots has become an indispensable means for real-time monitoring,dynamic tracking and post-disaster assessment of forest fires.In view of the shortcomings of the current fire point monitoring algorithm based on single remote sensing data and the low spatial and temporal resolution of fire point products,low algorithm accuracy,and poor adaptability,this paper combines multi-source remote sensing data to carry out high spatio-temporal resolution fire point monitoring and high-precision overfire Area estimation study.This article mainly studies fire point monitoring and fire area estimation from the following aspects:(1)In order to solve the problem that the fire point information is limited in space or time due to the uniqueness of the data in the remote sensing fire point monitoring,the optimal layout of multi-source remote sensing data in time and space is integrated to monitor fire point.First,use Himawari-8 / AHI data for real-time monitoring on the time series,and then use EOS / MODIS data,NPP / VIIRS,NOAA / AVHRR data,and FY-3 / VIRR data to perform time-interpolation analysis.This method is not only a fire point The highest temporal resolution of monitoring reached 10 minutes,and the highest spatial resolution reached 375 meters.(2)Analyze the minimum detectable fire spot area in the fire spot monitored by remote sensing satellite.According to the relationship between the wavelength and temperature in Planck's black body radiation law,using the different sensitivity of midinfrared band MIR and thermal infrared band TIR to high temperature,use The Newton iteration method solves the relationship between temperature and sensitivity in mixed pixels.The results show that between 300K-400 K,the sensitivity is increased from 90% to 10%,and between 400K-1500 K,the sensitivity is increased from 10% to 1%.(3)Considering the robustness and applicability of the fire point monitoring algorithm,an improved fire point monitoring algorithm suitable for a variety of remote sensing data is proposed.This algorithm performs threshold adjustments when the image is divided into day and night Optimized,added indexes such as NDVI,NDWI and NFPI,and combined the space and time information to conduct a confidence analysis on the fire point.Compared with 14.85% of Himawari-8 WLF,42.57% of NPP / VIIRS Active Fire and 52.48% of MODIS Active Fire,the algorithm improves the accuracy of fire point monitoring to 81.30%.(4)Take the fire in Xichang City as an example,using change detection and objectoriented analysis to extract the fire area before and after the fire,and obtain the fire area of 3049.6 hectares,combined with medium and high resolutions such as Landsat-8,Sentinel-2 and GF-2 The data analyzed multiple fires,and compared the results of GF-2,the overall accuracy of Landsat-8 and Sentinel-2 reached 91.99% and 92.14%,the error rate was 2.51% and 5.36%,and the leak rate was 29.04% and 25.87%,indicating that the use of Landsat-8 and Sentinel-2 has certain practical value for fire area estimation,but for small-scale fire analysis,GF-2's 1 square meter ultra-high resolution has great advantages.