Study On Signal Simulating And Processing Of Distributed Spaceborne Interferometric SAR System

Distributed spaceborne InSAR system is a novel spaceborne radar system combining satellite formation flying technology and spaceborne SAR technology. It can measure the ground elevation by formation flying and cooperation of several satellites. Because distributed spaceborne InSAR system has its own unique virtue relative to other ground elevation mapping technique and other InSAR system, it is becoming a hot point researched around the world. Aiming at this novel research, this paper studied system modeling, signal simulation, signal processing and DEM reconstruction of the distributed spaceborne InSAR system. The research can support realization of the system in the future.System modeling and signal simulaiton were studied in chapter 2. System models include geometric model and signal model. In geometric model, orbit model of the satellites, relative movement of the formation and geometric relation between satellite and ground were studied. Relative to traditional InSAR system the distributed spaceborne InSAR has the characters such as tridimensional baseline, variational baseline, squint-looking, separate transmitter and receiver. According to these characters correlation of signal was analyzed. In order to simulate the backscattered signal, backscattering model and facet model were studied. Using these models the electromagnetic scattering of the tridimensional ground scene was modeled. Then backscattered signal and SAR image were simulated. On the synchronization problem in the system, simulating method of synchronization error was studied. In order to realize surface decorrelation and volume decorrelation at the same time in simulated signal, the ground scene was separated into voxels based on facets. The method of separating voxels and facets was analyzed.Imaging and coregistration were studied in chapter 3. In some formations, the system is bistatic. In this situation, bistatic imaging is needed. At first a bi-velocity beeline distance model of bistatic imaging was presented. When error existed in the distance model, influence to imaging was analyzed. Based on this distance model, a improved Range-Doppler algorithm was studied. Then synchronization influence on imaging was analyzed. In the research of image coregistration, the emphases is sub-pixel coregistraion. In this dissertion two sub-pixel coregistration methods were studied: correlation coregistration and coregistration using spectral diversity. In correlation coregistration a method using chirp transform algorithm to increase computing efficiency was presented. In coregistration using spectral diversity an analysis from the point of view of statistical character was presented.Filtering and unwrapping were studied in chapter 4. Filtering of InSAR signal includes pre-filtering and interferogram filtering. Principle and method of pre-filtering were studied. In interferogram filtering, choice of filtering window size was analyzed. Estimating frequency parameters of interferogram is needed in pre-filtering and interferogram filtering. So the estimation using max-likelihood method was studied. Unwrapping of interferogram includes single baseline unwrapping and multi-baseline unwrapping. In single baseline unwrapping, path-following methods and minimum-norm methods were introduced. In multi-baseline unwrapping, based on geometric model of distributed spaceborne InSAR, the relation between several interferograms got from reconstruction formulas was presented. This relation was used in two multi-baseline unwrapping methods: max-likelihood method and least-squares method.DEM reconstruction was studied in chapter 5. In this chapter, three methods were studied: direct reconstruction, reconstruction with separate elevation and position measuring, reconstruction using Doppler formula of slave image. To these three methods the transferring coefficients between inputting error and reconstructing error were analyzed. The analysis shows that the requirement of baseline measurement is critical. Therefore a reconstruction method using ground control points was presented. The method can decrease the baseline measuring requirement. To the situation using single ground control point, transferring coefficient between inputting error and reconstructing error was analyzed.