Dem Generation And Accuracy Evaluation On Cross Interferometry With ERS-2/Envisat

As a new remote sensing technology, synthetic aperture radar interferometry InSAR firstly used in the1990’s has been successfully applied in the field of topographic mapping, surface deformation, disaster monitoring and so on. Recently, the increase of high quality SAR image data, e.g. ESA ENVISAT (in the early2000), greatly promote the development of the technique. As a common way to generate digital elevation model (DEM) products, the InSAR technique has the limitations, i.e. the temporal decorrelation and spatial decorrelation factors, for the production of a high-precision DEM using InSAR images. Therefore, this study proposes a Cross-interferometric SAR (CInSAR) method and real-life datasets has been used to validate the method.In this thesis, firstly the theories and principles of the Radar system (including RAR and SAR) and SAR interferometry (InSAR) are briefly introduced and the process of InSAR processing has been described which includes the influence of each parameter in the model for result’s accuracy, the description of primary interferometric phase components as well as the limitations in spaceborne InSAR DEM generation.CInSAR dramatically overcomes the baseline constraint by the slight difference of radar carrier frequencies, and characterized by a long baseline and a short repeat-pass interval.The CInSAR method will be introduced and the methods for the the critical step image registration, in the process of CInSAR processing will be proposed in order to resolve the problem of the traditional alignment method due to the long baseline and the severe topographic relief. Real-life data has been used in this study to validate the methods.In this thesis, an improved external DEM-assisted method is developed and employed to extract the DEM information in Alaska far north by C-band ERS-2and ENVISAT images. Finally, comparing between the generated results with the external NED DEM data as well as the accuracy analysis for the CInSAR-derived DEM, and modeling the orbit error containing in the DEM differences.