Study On DS-InSAR Method For Monitoring Time-series Surface Deformation In Mining Area

Coal resource is an important part of China's energy structure.Due to massive exploitation of coal resources,the area of surface subsidence in China has been increased year by year,and ecological environments have been damaged,and quality of cultivated land has been gradually reduced,and the phenomenon of soil erosion has been increasingly serious,and people's life and property safety are in threatening situations.Therefore,it is of great research value and practical significance to establish early warning system for surface deformation in mining area.The microwave sensor on the SAR satellite,which provides SAR images with different resolutions for scholars,has ability to work with all day and all-weather,and promotes the development of InSAR.At present,time-series InSAR method has been the research frontier of surface deformation monitoring.However,due to high vegetation coverage and large-scale subsidence,interference pairs of mining area are seriously affected by incoherence.Therefore,time-series InSAR method acquires point targets with low density in monitoring surface deformation of mining areas so that it is unable to effectively analyze the temporal and spatial evolution characters of surface deformation.In view of above problems,this paper studies distributed scatterer InSAR(DSInSAR)for deformation monitoring.Combining with deformation characters of mining area,this paper proposes two improved methods,one is homogeneous pixel selection with adaptive window,and the other is monitoring large-scale deformation of mining area by integrating probability integral model(PIM)and DS-InSAR.The proposed methods enrich the way of surface deformation monitoring and improve the accuracy of monitoring results.The main research work and results are as follows:(1)Main principles of DS-InSAR method are described.Sentinel-1 images,covering 94101 working face of Zhangshuanglou mining area from November 2018 to November 2019,are used as data sources.PS-InSAR and DS-InSAR methods are respectively used to monitor time-series surface deformation.The research results show that:(1)DS-InSAR method significantly improves the point density in the mining area,and obtains the complete range of surface subsidence,and provides the basis for analyzing temporal and spatial evolution characters of surface deformation in mining area;(2)By comparing and analyzing correlations between monitoring results of PSInSAR and DS-InSAR methods,this paper confirms the feasibility of DS-InSAR method for monitoring surface deformation in mining area;(3)Time-series monitoring results are verified by leveling data,and results show that DS-InSAR method based on SBAS data processing flow is less affected by spatial incoherence,and the accuracy of deformation monitoring results is higher than that of DS-InSAR based on PS-InSAR data processing flow.(2)In DS-InSAR method,window for homogeneous pixel selection is a fixed value,and the quality of interferogram after phase optimization will be affected by too large or too small window so that one adaptive window method for homogeneous pixel selection is proposed.Based on the average differential interferogram of interference pairs,the threshold of phase difference is set,and the optimal window for homomorphic pixel selection of each pixel is calculated.TerraSAR images,covering 9308 working face in Nantun mine of Yankuang grop from December 2011 to April 2012,are used as data sources.The homogeneous pixel selection and phase optimization are carried out respectively with adaptive and fixed windows.Research results show that the quality of differential interferograms and point density obtained by the former are higher than that obtained by the latter,which shows the effectiveness of the proposed method.(3)It is a problem that the magnitude of surface deformation in mining area exceeds the monitoring capability of time-series InSAR.In this paper,the probability integration model(PIM)for mining deformation prediction is introduced,and one method for large-scale deformation monitoring by combining PIM and DS-InSAR is proposed.The main idea is predicting the accumulated subsidence of mining area with geological mining conditions and parameters of working faces,and obtaining the simulated deformation phase of interference pairs in corresponding time period.In order to reduce the difficulty of phase unwrapping,residual phases of interference pairs are obtained by subtracting simulated phases from real phases.Taking TerraSAR images covering Nantun mine as the data source,the proposed method in this paper is realized.Research results show that the improved method can detect a greater level of surface deformation than DS-InSAR.Compared with leveling data,the accuracy of surface deformation extracted by this method is centimeter level,which is better than DS-InSAR monitoring results.There are 50 figures,11 tables and 88 references.