Study on air infiltration/escape area of underground coal fires using multi-source remote sensing data

Based on the changes of surface subsidence, temperature and vegetation caused by underground coal fires, the dynamic change process of fire area can be reflected to a certain extent.In this paper, the universal single channel algorithm and pixel dichotomy are used to retrieve the surface temperature and vegetation coverage in the fire area, and the differential interference “two track method” is used to extract the settlement characteristics.Combined with the distribution characteristics of surface subsidence, surface high/low temperature area and high/low vegetation coverage area, this paper attempts to analyze the high-temperature smoke escape area and low-temperature air entry area in the fire area.The results show that the high temperature area accounted for the largest proportion(27.18%) in 2013 and the low temperature area accounted for the largest proportion(16.51%) in 2015.The study area is mainly covered by class Ⅰ and class Ⅱ vegetation.The area of class Ⅰ vegetation is the largest in 2003(24.200 km², accounting for 99.48%).The high settlement value fluctuates from 9.591 mm to -3.124 mm, and the low settlement value fluctuates from -19.132 mm to -26.059 mm.From 2014 to 2017, the number of effective fracture points in the superposition treatment of temperature abnormal areas and class Ⅰ vegetation covered areas decreased from 5 617 to 863, and then increased to 3 578, with an overall decrease of 2 039.The number of effective fracture points in the superposition treatment of low-temperature area and grade Ⅳ vegetation coverage area increased from 4 to 13, an overall increase of about 3 times.The lowest settlement values of typical air inlet channel and typical flue gas escape channel appear in 20160131-20160319 interference pair, which are -262.402 mm and -232.832 mm respectively, while the highest settlement values appear in 20160506-20160530 interference pair, which are 602.910 mm and 564.202 mm respectively.