Data Processing And Simulation Of Interferometric Synthetic Aperture Radar

Interferometric synthetic aperture radar (InSAR) can provide accurate digital three-dimensional topographic measurement in all weather conditions day and night. It has been a hot topic in the area of synthetic aperture radar (SAR) in recent years. In 2000, the study on InSAR surged up worldwide due to the successful shuttle radar topography mission (SRTM) developed by NASA etc. Now, efficient and accurate data processing and simulation of InSAR are still the main problems in this area. In this dissertation, we try to study thoroughly the key techniques of the InSAR data processing and make an improvement on them in order for a higher efficiency, a less error and a more practicability. On the other hand, we present a novel approach to solve one of the intrinsic problems of InSAR simulation by a comprehensive comparison between the InSAR simulator and the conventional SAR simulator, which further deepen the understanding about the InSAR imaging.This dissertation is organized to first provide an overview of the InSAR theory, all kinds of decorrelation factors, and the error sources of the digital elevation model (DEM). Then, it is followed by a more detailed discussions on the main steps of the InSAR data processing flow and InSAR data simulation, including the complex image coregistration, flat effect of interferogram, noise filtering of interferometric image, phase unwrapping, the reconstruction of DEM, and so on.This dissertation is devoted to:1. In the discussion on the complex image coregistration, we grasp the commonnessand difference of the typical coregistration algorithms theoretically and practically.Then, we apply the quality factor concept in phase unwrapping to coregistration and present several potential evaluation functions that might be used in the coregistration operation, which further developed the evaluation-function-based coregistration algorithms of InSAR complex images.2. As for the noise filtering of interferometric image, according to the sources and statistic distribution of interferometric phase, we present an improved average filter and adaptive filter, which are proven to be effective both theoretically and practically.3. We discuss the representative phase unwrapping algorithms in detail, and make a comprehensive comparison among them by a series of unwrapping experiments with test and real interferometric data. Due to the difficulty of phase unwrapping, we probe into the evaluation of unwrapping results with different algorithms, and present an improved approach in which the intensity image, coherence map and other quality map is used synthetically to select the optimal unwrapping result.4. We analyze thoroughly on the representative InSAR simulators. With a clear understanding of the differences and commonness of the simulation on different levels, we present a new approach to compute the backscattering of the extended scene by the spectrum of complex images. The experiments have shown that this method effectively simulates the intrinsic baseline decorrelation effect of InSAR.