Monitoring The Surface Deformation Of Hongqinghe Coal Mine With Sentinel Satellite SAR Image

Underground coal mining has caused the increasingly prominent problem of surface subsidence,which poses a serious threat to personal safety and facility stability.It is necessary to monitor ? predict and early warning the surface subsidence continuously and comprehensively.Although the methods of leveling and GPS measurement have high accuracy,they have many problems such as heavy workload,low efficiency and sparse observation points,which make it difficult to describe the temporal evolution and spatial distribution characteristics of surface subsidence in detail.Synthetic aperture radar interferometry(InSAR)has become one of the commonly used surface subsidence monitoring techniques due to its advantages of wide coverage,high resolution,all-weather and non-contact.Especially,the time-series InSAR based on coherent targets(such as permanent scatterer PS,distributed scatterer DS,etc.),it is possible to perform deformation monitoring with mm-level accuracy.The Sentinel satellite which recently launched was able to make rapid repeated measurements in a six-day revisit cycle,providing a research conditions for studying the rules of surface subsidence caused by coal mining.This article is based on sentinel 1A satellite imagery,taking Hongqinghe Coal Mine as the research object,researches the time series InSAR monitoring method and the space-time evolution law of surface subsidence based on Sentinel satellite imagery.The main contents are as follows:(1)Optimized the registration method of Sentinel satellite imagery.The unique imaging method of Sentinel satellite makes the registration of images more difficult,enhanced spectral diversity is required on the basis of geometric registration to meet its registration accuracy.The conventional enhanced spectral diversity processing process uses end multi-view.When the theoretical coherence is less than 0.2,it is difficult to obtain an interference map that satisfies the registration accuracy.This paper optimizes the process of enhanced spectral diversity processing and adopts multi-view processing in advance which to further improve the registration accuracy and reliability and to ensure that the registration accuracy of the image meets the requirements.Through simulation experiments,it is proved that the improved processing flow has certain advantages,the algorithm is implemented by the algorithm,and the surface subsidence monitoring of Hongqinghe Coal Mine is successfully applied.(2)Introduced the processing flow of DS-InSAR,elaborated the key steps such as DSC identification,compared and analyzed the deformation monitoring capability and advantages and disadvantages of PS-InSAR and DS-InSAR,using sentinel 1A data and TanDEM,using PSInSAR and DSInSAR The technology extracted the surface subsidence information of Hongqinghe Coal Mine during 2017-2018,and found that the land subsidence area is about 32.3km2.Combined with the mining progress,the time evolution and spatial distribution of surface subsidence are analyzed,which clarifies that coal mining is the cause.The main cause of land subsidence.In addition,DS-InSAR is more suitable for this area than PS-InSAR.(3)Conducted research on surface subsidence prediction.Based on the theory of surface movement and deformation of coal mining,the main section is taken as the research object,and the subsidence prediction function is established by using support vector regression(SVR)model.According to the timing subsidence value extracted by DS-InSAR,the particle swarm optimization(PSO)algorithm is used to solve the relevant parameters,and the prediction of surface subsidence is realized.