Identification And Monitoring Of Underground Coal Seam Fire Area In Xinjiang Based On TIRS And InSAR

The burning of underground coal fire not only has an important impact on the national economy and the energy strategic security in China,but also has a very serious damage to the environment,ecology,vegetation and water resources.Xinjiang is one of the most serious coal field fires in the world.Although significant results have been achieved after decades of fire fighting and control,the scale of uncontrolled areas has gradually expanded and new fire areas have continued to appear.At the same time,due to the wide area and complex geological conditions,the conventional field investigation,geophysical exploration,geochemical exploration,drilling and other methods are often difficult to detect the wide-area coalfield fire area in Xinjiang.With its unique advantages,remote sensing detection methods make up for the shortcomings of conventional methods to detect coalfield fire areas.At present,the research of using thermal infrared remote sensing to detect fire area has been more mature,and the research of using InSAR technology to monitor the surface deformation of fire area has also begun to attract attention.However,a single remote sensing detection method is often difficult to meet the actual needs of accurate detection of wide area coalfield fire areas in Xinjiang.In order to identify and monitor the coal field fire area in Xinjiang more accurately and reliably,a method of identifying and monitoring the fire area of coal field in the wide area is studied by using TIRS and InSAR technology and applied in the Zhundong coalfield in the north of Xinjiang.The main contents and achievements of the study are as follows:(1)Taking the two Landsat 8 satellite images covering the study area imaged in summer and winter as the research data,and taking the eastern part of Junggar coalfield in Xinjiang as the study area.The method of the radiative transfer equation,the monowindow algorithm,the single-channel method and the split-window algorithm are used to carry out comparative experiments on the inversion of land surface temperature.The results of these four temperature inversion algorithms are verified and analyzed using the land surface temperature product MOD11A1 of MODIS.The experimental results show that the mono-window algorithm is the most suitable land surface temperature inversion algorithm for this study area.Finally,the mono-window algorithm is used to obtain the land surface temperature of the study area from April 2017 to March 2018.(2)Most of the existing researches on coal fire detection based on thermal infrared remote sensing are limited to the extraction of land surface temperature anomaly in a small range,which is not suitable for the detection of wide area coal fire in Xinjiang.In order to solve this problem,a method of temperature anomaly extraction based on super-pixel segmentation is proposed,which innovatively applies the super-pixel segmentation method in the field of image segmentation to the extraction of land surface temperature anomaly in coalfield fire area.The experimental results show that compared with the traditional threshold segmentation method,this method obviously eliminates many sporadic and isolated temperature abnormal points,and greatly improves the accuracy of wide area coalfield fire detection based on thermal infrared remote sensing.(3)Due to the complex geological conditions and the lack of permanent scatterers such as rocks and artificial buildings,it is difficult to obtain enough deformation monitoring points on the surface of coal fire area by traditional time series InSAR methods.In order to solve this problem,the basic principle of DS-InSAR is studied,and the DS-InSAR method,which combines the fast homogeneous point selection algorithm and the eigendecomposition method of covariance/coherence matrix,is used to monitor the surface deformation of the study area from April 2017 to March 2018.The experimental results show that the DS-InSAR method has great advantages and potential for dealing with coal field fire areas that lack permanent scatterers.It greatly increases the density of the surface deformation monitoring points,and can more truly reflect the surface deformation of the fire area.(4)Aiming at the problem of how to identify and monitor the wide area coalfield fire area in Xinjiang by combining the land surface temperature anomaly and deformation information,an overlay analysis method based on surface temperature anomaly,surface deformation and other remote sensing information is proposed.In this method,the temperature anomaly area including surface deformation points is regarded as the possible coal fire area.Moreover,according to the characteristics that the heat released by the combustion of underground coal fire will lead to the melting of surface snow,some non-coal fire areas are excluded.Finally,combined with some field data and satellite image data,the location of Jiangjun Gebi fire area is identified.The experimental results are consistent with the results of the fourth survey of fire area in Xinjiang by Xinjiang coal field fire extinguishing Engineering Bureau,which verifies the accuracy of identifying fire zone by combining TIRS and DS-InSAR.(5)Taking the Jiangjun Gebi fire area as the study area,using the method of overlay analysis,the long-term dynamic monitoring of Jiangjun Gebi fire area can be realized by obtaining the range of Jiangjun Gebi fire area at the end of 2018,2019 and2020.According to the range of fire area and the change of land surface temperature,the combustion state,development trend and treatment results of fire area are analyzed.Finally,according to the characteristics of time-series accumulated deformation of the surface in the fire area and the study of the correlation between the surface deformation rate and the temperature anomaly,the feasibility of identifying and monitoring the Jiangjun Gebi fire area based on TIRS and DS-InSAR is verified.There are 41 figures,6 tables and 108 references.