Identification And Monitoring Analysis Of Xinjiang Coalfield Fire Area Based On Multi-Source Remote Sensing

Coal fire has caused serious negative impact on ecological environment,infrastructure and safety of coal mine.There are many outcrop coal seams,the thickness of coal seams,the shallow burial and the short combustion period,the coalfield fire area in Xinjiang has become one of the most serious fire areas in our country and even in the world.Although good results have been achieved after many years of fire control,the burning scale of uncontrolled fire areas has gradually increased,and many new fire areas have been formed.Therefore,it is of vital significance to identify and monitor the coal fire area in Xinjiang timely,accurately and effectively.However,traditional detection methods are often difficult to meet the actual needs due to the problems such as high risk,high cost and small scope.In contrast,remote sensing technology makes up for the shortcomings of traditional methods with its unique advantages.However,there are still many problems in the detection of coal fire areas by existing remote sensing technology,which needs to be further studied.Therefore,this thesis studies a refined identification method of coalfield fire area based on multi-source remote sensing.Taking the Jiangjungebi fire area and Jiangjunmiao fire area in Zhundong coal field of Xinjiang as the research object,the fine identification of the fire area is carried out,and the multi-source remote sensing is used to carry out long time series monitoring and analysis of the identified fire area.The main research contents and achievements are as follows:(1)The basic theory of land surface temperature inversion is expounded,and the single window algorithm is used to invert the land surface temperature in the study area according to the existing research.Based on the basic theory of InSAR technology,the principles of D-InSAR and time series SBAS method are introduced.On this basis,distributed scatterers,homogeneous point selection algorithm(KS test,BWS test,GLR test and HTCI algorithm)of DS-InSAR method and phase optimization method of coherence matrix eigenvalue decomposition are introduced in detail.Finally,the principle of multi-source remote sensing data comprehensive identification of coal fire area is summarized and introduced.(2)A surface deformation monitoring method of coalfield fire area based on DSInSAR is studied.Firstly,KS test,BWS test,GLR test and HTCI algorithm are used to select homogeneous point in the study area,and the results are compared and analyzed.It is found that HTCI algorithm is more suitable for the identification of the same particle in the fire area.Then,the coherence matrix eigenvalue decomposition method is used to optimize the phase,and the average coherence of the fire area is improved from 0.67 to 0.81.Finally,the surface deformation in the study area was calculated.Compared with the traditional time-series SBAS method,the monitoring point density of DS-InSAR increased by 2.74 times,which can provide more comprehensive and detailed surface deformation information for the coal fire area.(3)A thermal anomaly information extraction method based on temperature gradient adaptive threshold is proposed.This method combines the spatial characteristics and mathematical statistical characteristics of thermal anomalies caused by coal fire combustion,and can adaptively calculate the corresponding temperature threshold for different fire areas.Meanwhile,it can also effectively eliminate the thermal anomaly areas caused by non-coal fire factors.The thermal infrared remote sensing data of the study area is used to verify the effectiveness of the proposed method.(4)A refined identification method of coalfield fire area based on multi-source remote sensing is studied.Firstly,land surface temperature is retrieved by Landsat-8thermal infrared image,and thermal anomaly information extraction method based on temperature gradient adaptive threshold is proposed.Then the Sentinel-1 image and DS-InSAR method were used to calculate the surface deformation information.The suspected coal fire area is obtained by overlay them.Finally,NBR of Sentinel-2 image inversion and high definition image comprehensive analysis were used to determine the combustion type of suspected coal fire area,so as to achieve fine identification of coal fire area.The multi-source remote sensing data from April 2017 to October 2018 were used to identify the fire area in the study area,and the measured data were used to verify the analysis.The results show that the accuracy of coal fire area identification is 82.9%,and the coal fire combustion type can be effectively identified,which proves the reliability and practicability of the method proposed in this thesis.(5)A long time series monitoring and analysis of coalfield fire areas based on multi-source remote sensing is studied.The information of multi-source remote sensing monitoring during April 2017 to January 2022 is used to monitor and analyze the identified fire areas.Thermal infrared data from 2001 to 2021 is used to monitor the coal fire burning trend.The results show that:(1)The spread trend of coal fire combustion is closely related to the direction of coal seam combustion;(2)The burning intensity of open air coal fires has obvious seasonal characteristics.The burning intensity of open air coal fires is more intense in autumn and winter,but it decreases somewhat during snowfall due to winter snowfall.(3)There are serious surface subsidence phenomena in coal fire area,coal fire combustion and dump subsidence are the main causes of fire area subsidence,and different types of subsidence areas have different subsidence magnitude and sequential subsidence trend.(4)There is a strong coupling and negative correlation between surface temperature and subsidence,which proves that it is feasible to identify coal fire area by superposition of surface subsidence and thermal anomaly information.There are 41 figures,tables 6 and 124 references.