Monitoring Of Surface Deformation With Time-series InSAR Technique Based On Multi-source SAR Data

Differential synthetic aperture radar interferometry (D-InSAR) is a new space-to-earthobservation technology which uses the phase information carried by SAR images to obtainthe surface deformation with high precision. Because of its high resolution, high precision andwide coverage, it has been widely applied in the monitoring of surface deformation such asland subsidence. With the successful launch of ERS-1/2, ENVISAT ASAR, ALOS PALSAR,TERRASAR-X, RADARSAT-2and COSMO-SkyMed SAR satellites, the SAR data ofdifferent wavelengths and different orbits can cover the same area and have a certain overlapin their observed times, so research on the integration of multi-source SAR data not onlybecomes true but also becomes the necessary scientific subject. Based on this, using the SARdata of different senors, different tracks to obtain the monitoring results with shorter periodand denser time series will provide the more reasonable and scientific basis for the preventionand control of the geological hazards such as ground subsidence.However, conventional differential synthetic aperture radar interferometry is oftenlimited by the temporal and spatial baseline decorrelation, DEM error and atmospheric delaywhich lead to hardly obtain the reliable monitoring result. In addition, for the same researcharea, the deformation monitored by different SAR sensors is often different and the singleSAR sensor has its specific revisit period, relying on the SAR data of a single orbit can onlyobtain relatively sparse time-series results, which do not reflect the surface deformation of theresearch area accurately. But the integration of multi-source SAR data can realize theunification of different data and also obtain denser time-series result.To solve the above two problems, taking Taiyuan as the research area, the paper studiesthe time-series InSAR technology of the points with stable phase in the SAR images and therelated theories and experiment analysis of multi-source SAR data in order to obtain thesurface deformation with high precision. Accordingly, the main researches and achievementsin the paper are as follows:(1) Introduce the basic principles of InSAR and D-InSAR technique and present the errorsources which affect the precision of D-InSAR in the deformation monitoring and thecorresponding reduced measures systematically. Then, sum up the principles of the currentlypopular new D-InSAR methods based on the targets with stable phase information.(2) As to each error source, make a detail study of the Interferometric Point TargetAnalysis, including the target recognition methods in the time-series SAR images, such as based on the amplitude and intensity, the coherence coefficient and the frequencycharacteristics of SAR images and point out its applicable conditions. Then, use thistechnique to monitor the land subsidence in Taiyuan city. Firstly, select the high coherenttargets with stable scattering characteristics by combining the amplitude dispersion indexthreshold and point target detection method with the obtained ENVISAT ASAR data andusing the “SBAS+PS” and single “PS” to generate the interferograms. Secondly, use thetwo-dimensional phase linear regression analysis model to analyze the phase observationsiteratively to separate the DEM error and linear deformation rate and filter the residual phasein time and spatial domain according to its different features to seperate the atmospheric delayand orbit error with obvious trend and obtain the land subsidence results. Thirdly, through thecomparision of these two results, we find that “SBAS+PS” can obtain results withmillimeter-level precision even in fewer amounts of data.The paper also uses the single “PS”to process the TERRASAR-X data which has a very high resolution. Through thecomparision of ENVISAT ASAR data, it has a high consistency of the result obtained byENVISAT ASAR.(3) This paper studies the basic principles and key technique in the data processing ofmulti-source SAR data integration and obtains the surface deformation information withshorter period and denser time-series results in Taiyuan city by using the C-band ENVISATASAR and X-band TERRASAR-X images which reflects the surface deformation of Taiyuanmore realistically.(4) This paper obtains the monitoring results from1992to2013by using the SARimages with different wavelengths, different sensors and different resolutions and analyzesthe distribution of the land subsidence at each stages, having a comprehensive knowledge ofthe developmental history and status quo of land subsidence in Taiyuan city, in order toprovide basic information to the study of its mechanism and trend analysis. The comparion ofthe InSAR results with the leveling results and GPS results at the same monitoring periodproves high consistency.