Study On Atmospheric Delay Correction Of InSAR Time Series To Detect Ground Subsidence In Mining Area

Coal resource is one of indispensable and basic energy sources for social and economic development,while long-term exploitation of underground coal often leads to ecological and environment hazards such as ground fissures,soil degradation and ground surface subsidence.So it is of high value to effectively monitor the ground subsicence.In the papar,we acquire 44 sences of dual-polarization(VH and VV)Sentinel-1(S1)images to monitor ground subsidence of Handan from Jun.2015 to Aug.2017,combining with In SAR time series technology and two atmospheric correction models of PL and GACOS(Generic Atmospheric Correction Online Service for In SAR,GACOS).Then we analyze the accuracy of monitoring subsidence by basing on the leveling measurement data,focus on the subsidence results of the main regions in Fengfeng mining area,and discusse the issues related to atmospheric delay correction.In addition,we obtain the information of subsidence areas and their contours.The main results and conclusions in this thesis are as follows:(1)It is systematically summarized the principle of atmospheric correction methods in the In SAR technology of removing atmospheric delay phase,whose advantages and disadvantages are also compared with;the two proposed atmospheric correction models combined with In SAR time series,such as the phase based Powerlaw(PL)and external data based GACOS,are introduced in detail.The comparative analysis shows that by taking into account the spatial variation of the troposphere and relationship with topographic factors,the PL model is less affected by the time factor;the GACOS model has the advantages of wide coverage,high spatial resolution,and near real-time.(2)According to the characteristics of S1 data in azimuth registration,it is also summarized the specific processing flow.The monitoring subsidence results of the two data sets covering Handan City are obtained through PS-In SAR.The statistical analysis shows that the difference including: under the same conditions of images processing,the number of PS points selected by the VV data set is about 16% higher than that of the VH data,and most of them are distributed near the mountainous terrain where the terrain is undulating;from the intensity information,it is found that the average backscattering signal of the VV dataset is stronger.While the same results include that: the coherence of two datasets is basically the same by the time;the two kinds of subsidence results are well consistent with.The above two results together indicate that the significant subsidence areas during the study period are distributed in the central and northern parts of Chengan County and Feixiang County,especially at the junction of Feixiang County,Yongnian County,and Quzhou County,where the maximum annual average subsidence rate exceeded-50 mm/year,while the main influencing factor is the long-term exploitation of groundwater resources.(3)After sub-block processing,we acquire the sub-image set mainly covering the Fengfeng mining area.By the processing of In SAR time seires again,the subsidence results removing atmospheric delay effect are obtained by two kind models of PL and GACOS.The analysis also shows that for the same atmospheric correction model,the phase delay results of the two datasets are consistent,while the average delay phase difference does not exceed 3 rad.But the distribution of delay phase reflects a spatialtemporal heterogeneity.Then by combining with the measured data,the monitoring results of same name points(SNPs)is extracted,from which the statistical analysis found that the atmospheric delay of the VV data set is relatively stable with PL mode.And the correlation between the SNPs and the level monitoring results is the highest,as the coefficient is 0.9728.(4)After PL mode,we combine the two kinds of results removing atmospheric delay in the Fengfeng area,perform Kriging interpolation for each period of monitoring results and draw the time-sequence contour maps of the main mining subsidence in the two periods from May 2016 to Dec.2016 and Mar.2017 to Jul.2017.The above results show that the cumulative subsidence of the PS points in the selected area is improved by the cm level.According to the spatial distribution of subsidence contours and the changing trend over time,four subsidence areas are detected in the main mining area.Then,taking the Wutongzhaung Mine as an example,by comparing the results before and after atmospheric correction with the coal mining information,found that the results after atmospheric correction have high feasibility in detecting the subsidence area and its boundary.