Subsidence Monitoring Along High Speed Railway In Loess Area Based On DS-InSAR Method

The normal communication of high-speed railways is closely related to the safety of people's lives and properties,its extremely fast operating speed puts higher requirements on the stability of the roadbed and auxiliary facilities?e.g.,bridges,tunnels?.In recent years,with the vigorous development of China's high-speed rail industry,many high-speed rail lines cross the loess area of China.However,the engineering properties?water sensitivity,macroporosity,structural?of loess and human activities are highly prone to land subsidence,which poses a great threat to the safe operation of high-speed rail.Therefore,it is very important to monitor the subsidence along the high-speed railway in the loess area.However,the large span of high-speed railways and the complex geological environmental along the line make the traditional subsidence monitoring methods?such as leveling,GPS measurement,etc.?have obvious limitations.Therefore,it is necessary to use a monitoring method with a wide range of monitoring and time-saving and labor-saving.Differential synthetic aperture radar interferometry?DInSAR?,with some prominent advantages such as pantoscopic view,high ground resolution and high automation,is widely applied for monitoring and investigating land subsidence field.However,the subsidence measurements derived by DInSAR are often degraded by spatial and temporal decorrelation and atmospheric artifacts.In order to overcome the above difficulties,the Persistent Scatterer InSAR?PSI?developed in recent years has become a breakthrough in the realization of land subsidence monitoring,its basic principle is to first identify the permanent scatterers?PS?,and then separate the land subsidence information based on time-series differential interference phase with high SNR of permanent scatterers.However,the surface cover is dominated by sand in the loess area,and the number of available PS points is small,making it difficult to obtain the spatial details of land subsidence.It has been pointed out that in non-artificial surface areas?such as barren hills and deserts?,distributed satellites?DS?have value in monitoring surface subsidence.The DS point target is composed of adjacent target pixels with similar scattering characteristics in the SAR imaging,and its mechanism characteristics involve the coherent accumulation of all smaller scatterers in the element.The SqueeSAR^TM uses hypothesis testing as a means to select the homogenous points of the target pixel and perform adaptive multi-view,and then combines with PS and DS points to perform the deformation inversion.The results show that the density of DS points can be increased by at least 4 times compared to PSInSAR^TM.In order to improve the density of points and the details of land subsidence in the loess area,this paper uses DS-InSAR method to monitor the subsidence along the high-speed railway in the loess area,based on the multi-temporal C-band Sentinel-1A SAR images.Firstly,according to the nature of loess area,this paper focuses on the selection of interference pair and DS point in the DS-InSAR method,and extract the time series subsidence information of high-speed railway in the loess area.Finally,this paper evaluates the accuracy of the subsidence results.The details are as follows:?1?According to the network of interferograms and the type of interferometric decorrelation,this paper analyzes the sensonal coherence in the loess area,and concluded that the loess area has no seasonal decoherence characteristics based on the variation law of the coherence coefficient is obtained by the interference,the time baseline threshold is the most important factor in the interference selection in the loess area.Based on the above analysis,this paper analyzes the interference combination,which is mainly to select the time baseline threshold and increase the participation of winter SAR image interference pairs,so as to improve the overall coherence of the interference network.?2?In order to improve the accuracy of DS point selection in loess area,this paper compares the efficiency and adaptability of three SHP identification methods?GLRT,BWS,FaSHPS?,which is used to search the homogeneous pixels.The results show that the FaSHPS algorithm is more efficient and adaptive in SHP recognition in the loess ares.?3?We monitor ground subsidence by Sentinel-1A?acquired between 2017 and 2018?images in Zhengxi high-speed railway.The subsidence monitoring of the entire line of Zhengxi high-speed railway was carried out by three sets of interference pairs,and two obvious subsidence areas?Mianchi and Weinan?distributed in the high-speed railway line were detected.Analyzing the result of DS-InSAR,we find that subsidence rate of Mianchi and Weinan are-116.67mm/yr⁴1.88mm/yr and-50.92mm/yr^-1.3mm/yr.Compared with PS-InSAR method,DS-InSAR method can improve the reliability and effectiveness in the field of surface subsidence monitoring.