A New Displacement Estimation Algorithm Based On Sentinel-1 Consecutive Interferograms

Persistent Scatterer Interferometry(PSI)is newly developed remote sensing technique,this technique uses multi acquisition Synthetic Aperture Radar(SAR)images,organized them in a specific way to form numerous interfero grams,then select high coherence points which have stable phase value in interferograms,and finally extract surface displacement from these points.Since the first PSI approach introduced in 2000,due to this technique can detect Earth's surface displacement in large scale,long period,and high accuracy,it has been widely concerned and developed quickly.PSI now is suitable for many SAR satellites'(ERS1/2,ALOS,TerraSAR-X and so on)data processing,and has been widely used in many field like city subsidence monitoring,geological disaster assessment,and glacier drift measurement,plays a key role in national life and production.In this case,studying new PSI model and algorithm has become both important in theory research and realistic meaning.Sentinel-1 is European Space Agency's(ESA)newest SAR mission,since the launch of Sentinel-1A and Sentinel-1B in 2014 and 2016,it has been widely concerned for the data is free to download,have shorter revisit time,more precise orbit control and more advanced imaging mode.It has become a new hot research to apply this new data to PSI processing to achieve more precise and reliable results.In this paper,we make use of these advantages of Sentinel-1 data,propose a new PSI approach base on consecutive interfero grams,instead commonly used single master(PS approach)or small baseline subset(SBAS approach),this new approach use two images adjacent in acquisition time to form interfero gram,which means for each image,it is the master image of next acquisition image as well as the slave image of previous image,so that interferograms can be formed consecutively in time.This new approach makes use of the stable revisit time of Sentinel-1,solve topographic error phase independently by making difference between nearby interfero grams to cancel linear deformation,to improve the accuracy of estimation We use 32 descending Sentinel-1B images over the Tianj in City,acquisition time from Nov 2016 to Jan 2018,apply this new approach to detect subsidence,generate mean subsidence velocity map over this area,and finally validate the accuracy of this new approach by comparing to levelling data.The main work and contribution of this paper are:(1)Implement the consecutive interferograms approach.This new approach can keep PSs selection method and minimum the decorrelation caused by time to improve coherence compare to PS approach,and do not need to apply filtering or multilook ing,reduce interfero grams amounts and simplify workflow compare to SBAS approach.(2)Propose new displacement signal estimation algorithm.This new algorithm makes use of the stable revisit time of Sentinel-1 data,cancel the linear deformation by making difference between nearby interfero grams,solve topographic error phase independently,reduce the impact this term to displacement estimation,avoid spatial filltering,improve the accuracy of phase unwrapping and displacement estimation.(3)Provide new ideas to PSs selection and phase unwrapping algorithms.By forming an interfero gram between the last and the first image additionally,a "ring interferograms" can be formed,the difference between nearby pixels phase summation in ring interferograms is approximately integer times of 2?,the residual value can help to identify PSs points whereas the integer value can add a constraint on 3-D phase unwrapping.