Mapping of earth deformations with satellite SAR interferometry: A study of its accuracy and reliability performances

There are a wide range of methods for the measurement of earth deformations, among which the space-based satellite SAR interferometry (InSAR) possesses a number of advantages and has been increasingly used in recent years.; Despite its advantages and increasing popularity as well as a large number of demonstrated successful applications, InSAR however still has its shortcomings that often severely limit its applications. It is therefore of significance to both better understand and to overcome these shortcomings. This thesis is dedicated to this endeavour by focusing on the analysis and improvement of the two most important performance aspects of InSAR, its measurement accuracy and reliability.; Deformation measurements with InSAR can be degraded by errors in radar observations and in its system parameters, including primarily the errors in the topographic data in InSAR processing, the phase observations and the baseline parameters. Uniform error models are proposed for the three differential interferometric approaches with considering the various error sources. The models are found useful in making error budget estimations for different InSAR configurations. It is also found that the topographic data used in differential interferometric processing may cause either random or systematic errors in the InSAR results depending on the error nature of the topographic data. Gaussian noise in the topographic data causes random errors in the deformation measurements and reduces the coherence of a deformation interferogram while systematic errors in the topographic data generate systematic bias in the deformation measurements.; To improve the quality of interferometric phases, a method for the priori filtering of the SAR data is developed and tested. It is found that the method can effectively raise the SNR of interferograms and hence greatly benefiting the subsequent interferometric data processing. To refine baseline parameters, a new method that is based on the use of ground control points (GCPs) and least-squares estimation is proposed. The effects of the number, accuracy and distribution of GCPs on baseline estimations are examined with a series of simulation studies. Recommendations on effective use of GCPs are given based on the studies from a practical standpoint.; It is the first time that the concept of reliability is systematically introduced into InSAR data analysis. Besides looking into the effects of gross errors on InSAR measurements, two areas are specifically studied to enhance the reliability of InSAR systems, (1) the design of favourable GCP/tiepoint distribution configurations to provide stronger geometric strengths for gross error detection; and (2) robust estimation of the parameters in co-registration and baseline estimation. (Abstract shortened by UMI.)...