ABSTRACT:More and more high-speed railway are under construction in China. The slow settlement along high-speed railway tracks and newlybuilt stations would lead to inhomogeneous deformation of local area, and the accumulation may be a threat to the safe operation of high-speed rail system. In this paper, surface deformation of the newly-built high-speed railway station as well as the railway lines in Shenyang region will be retrieved by time series InSAR analysis using multi-orbit COSMO-SkyMed images. This paper focuses on the non-uniform subsidence caused by the changing of local environment along the railway. The accuracy of the settlement results can be verified by cross validation of the results obtained from two different orbits during the same period.* Corresponding author
INTRODUTIONIn recent years, China has ushered a new era with a lot of highspeed railway infrastructure construction, indicating the rapid economic development of the whole country. The pressure on the roadbed and elevated bridges of high-speed railway is much higher than traditional rail tracks because of the extremely high speed. The slow settlement along and around the high-speed rail tracks may also threaten the safe operation of high-speed trains after a long period of accumulation. Besides, newly-built highspeed railway stations may also lead to inhomogeneous deformation of the surrounding area. Therefore, continuously monitoring the settlement of high speed rail tracks and stations is a key to ensure the safety and long-term development of public transportation safety (Dai, 2013;Liu, 2016;Shi, 2015;Yang, 2015).
DEFORMATION MONITORING WITH TIME SERIES INSARTime-series InSAR utilizes stable points which are not influenced by saptial-temporal decorrelation and suppresses the atmospheric delaying phase. The evolution of surface deformation is obtained by separating the atmospheric phase and deformation phase in time series based on their different spatial-temporal characteristics. In our data processing, interferograms are firstly generated after selection of a common master, image coregistration and resampling of slave images. Secondly, the phase variation caused by flat earth with reference to the ellipsoid is estimated and removed, followed by geocoding of the flattened interferograms. Then, amplitude and phase analysis are conducted to estimate the probability of individual pixels as PS candidates, resulting in a subset of PS candidates. Next, the selected PS candidates are filtered to reject those that appear to be persistent accidentally which are dominated by scatterers in adjacent PS pixels or are only persistent in certain interferograms. For each PS point, the wrapped phases in differential interferograms can be decomposed into uncompensated topography, motion of the target between different acquisitions, the object scattering phase related to the path length traveled in the resolution cell, atmospheric phase delays, the phase caused by imprecise orbit data, and additive noise. Finally, the phases are unwrapped and the unwra...