Study On The Deformation Monitoring Based On Integrating Of Multi-source SAR Imagery

Interferometric Synthetic Aperture Radar(In SAR), a newly arisen space geodetic technique developed since 1960 s, has been proved as a powerful technique for producing topographic maps, and measurements of deformation caused by earthquake, landslide and those of over pumping underground water with high precision and high spatial resolution over large area. However, the In SAR may fail in deformation monitoring due to the large deformation gradient in time and space domain as the phase recorded in SAR system is wrapped. Besides, only one dimensional deformation along the line of sight(Lo S) can be detected using In SAR technique as the SAR images are acquired with side-looking mode. These limitations reduce the accuracy and hamper the practical application of In SAR.Fortunately, many of spaceborne SAR satellites has been launched recently, which allow us to observe the earth using SAR images with different wavelength, different repeat time, different spatial resolution and different incidence angle. Therefore, this thesis focuses on investigating the models and algorithm of integrating of multi-source SAR images acquired by different SAR satellites to increase the accuracy of In SAR and detect the three-dimensional deformation field. The main achievements and innovations of this dissertation are as follows:(1) The western Tianjin is selected as study area, where the ground subsidence due to ground water overpumping has been inversed by processing 14 scenes of Envisat ASAR(ASAR) images and 29 scenes of Terra SAR-X(TSX) images acquired during January 2009 to September 2010. Both of these results show that the study area suffers seriously from subsidence in this period. And the comparison between these results also shows that the X-band TSX data with high spatial resolution can be optimization selected by combining the C-band ASAR data with the advantage of large coverage. Besides, the consistency of subsidence detected by ASAR and TSX in common coherent points indicate that the accuracy of deformation can be confirmed by comparison of the results derived from different SAR images.(2) A methodology to integrate multi-source SAR images(e.g. ASAR and TSX) has been proposed to increase the accuracy and enhance the robustness of time series In SAR(TS-In SAR) analysis. The rationale of the methodology is that the deformation rate is firstly calculated based on time series ASAR images, which are incorporated and subtracted from TSX differential interferograms, and the residual deformation can be estimated more reliably based on these double-differentiated TSX images by TS-In SAR analysis. 14 ASAR images and 29 TSX images acquired between January 2009 and September 2010 over western of Tianjin area are used as test data. The results show that the methodology achieves a higher accuracy than that from the 29 TSX images alone. And the distribution of ground subsidence derived from the integrated process is more realistically.(3) A new algorithm used for deformation monitoring during long time interval based on DIn SAR has been proposed. The effectiveness of In SAR has been analysis and found that the phase variation is mainly limitation of DIn SAR technique and TS-In SAR technique for deformation monitoring. A new algorithm used for solving this limitation by integrating multi-source SAR images has been proposed. As most of deformation information, derived from low spatial resolution SAR images such as ASAR, has been subtracted from high spatial resolution SAR images, the phase variation can be reduced to meet the need of phase unwrapping. The deformation rate derived from DIn SAR using 2 TSX images and from TS-In SAR using only 11 of 29 TSX images is agree with leveling measurements based on this algorithm. The results show that this algorithm can be used for deformation monitoring with a few of high spatial resolution TSX images with long time interval. Besides, this is of great significance in practice because fewer budget would be spend in purchase of high resolution TSX data.(4) A new approach used for inversion of probability integral method parameters has been proposed based on 3-D deformation derived from In SAR. The three-dimensional(3-D) model and algorithm which integrating multi-source SAR images acquired with different geometric parameters in different satellite platform has been proposed. 3-D deformation, detected using two pairs of ASAR images acquired from two different tracks and one pair of Pal SAR images, has been proven by comparing with leveling measurements and used for inversion of probability integral method parameters. The subsidence caused by underground mining has been predicted using these parameters and compared with leveling measurements. The results show that this approach can be used to inverse the probability integral method parameters based on the 3-D deformation. And this approach overcome the limitation that the real ground deformation can’t be detected by sparse observation stations.