Application Research Of Time-Series InSAR Technology In Mining Subsidence Monitoring

Large-scale or over-exploitation of coal resources has led to a series of ecological environment,surface subsidence and people's safety problems in mining areas,and the increase of small coal mines has aggravated the seriousness of these problems.Therefore,it is urgent to carry out timely and efficient comprehensive monitoring and analysis of mining subsidence,so as to master the law of mining subsidence and provide a reliable basis for mining and governance of mining areas.As an advanced earth observation method,Interferometry Synthetic Aperture Radar(InSAR)technology has developed rapidly due to its all-weather,all-weather,and large-scale features,and has been widely used in surface mining subsidence monitoring in mining areas.Especially,good results are obtained in the monitoring of surface subsidence in small and slow mining areas.However,the mining area has the characteristics of large magnitude and fast speed,which can easily cause problems such as decoherence,atmospheric delay and low precision of DEM,so that the traditional InSAR technology is limited in the application of surface subsidence monitoring in mining areas.Aiming at this problem,this thesis adopts time-series InSAR technology to monitor mining subsidence in Huangling mining area of Shaanxi Province.The main research work and conclusions are as follows:(1)Taking Huangling mining area of Shaanxi Province as the study area,the 12-scene ALOS PALSAR(2007-2009)and 92-scene Sentinel-1A(2017-2020)data covering the study area were selected.Stacking-InSAR technique was used to detect the subsidence in two different periods,and the annual average subsidence rate of Huangling mining area was obtained successfully.Then,combined with the relevant data of working face mining in Huangling mining area,ArcGIS software was used to compare and analyze the above two monitoring results.The results show that the monitoring results of the Stacking-InSAR technology are consistent with the distribution of the working face in the mining area,which indicates that the Stacking-InSAR technology can be Monitoring of large-scale surface subsidence in the mining area,and the monitoring results are relatively reliable.(2)The Sentinel-1A data with 20-scene covering the full mining time of 803 working face were selected.Firstly,the combined model of Small Baseline Subsat and Interferometric Point Target Analysis technology(SBAS-IPTA)was used to obtain the annual average subsidence rate and time-series cumulative subsidence.Combined with the relevant mining data,the profile,contour and feature points of the monitoring results were analyzed.And it is concluded that the subsidence area is in good agreement with the distribution of working face and mining situation.Then,the 20-scene Sentinel-1A was processed by multi-main image coherence point target analysis(IPTA)technology,and the average annual subsidence rate and time-series cumulative subsidence were extracted.The monitoring results were compared and analyzed,and it was found that the monitored subsidence rate and maximum cumulative subsidence were about 15 mm different from the SBAS-IPTA monitoring results.And the subsidence area is consistent with the distribution of the working face,which indicates that the multi-master image IPTA technology can also be applied to the small and slow surface subsidence of the mining area.(3)The time-series cumulative subsidence monitored by SBAS-IPTA was combined with the Support Vector Regression(SVR)algorithm to predict mining subsidence.Firstly,the cumulative subsidence of time-series was divided into training set and test set.Then the training set was used to construct the model.Finally,three precision evaluation index parameters of mean absolute error,root mean square error and determination coefficient were used to evaluate the accuracy of the predicted results of the test set.The predicted results of SVR are consistent with the monitoring results of SBAS-IPTA.The results show that the mining subsidence prediction model constructed in this thesis has high accuracy,meets the engineering requirements of mining area,and provides a method for mining subsidence prediction.