Study On Polarimetric Synthetic Aperture Radar Interferometry And Tomography

Polarimetric synthetic aperture radar(SAR) interferometry integrates the perception of polarization and the traditional interference information and thus holds the all-time, all-weather, long-range capabilities of targets’ electromagnetic characteristics, spatial position, structure detection and environmental monitoring. It has shown great advantages and potential in applications such as topographic mapping, marine monitoring, agriculture and forestry, military reconnaissance, surveying and mapping weather graphics, environmental protection and disaster monitoring. Nowadays, polarimetric SAR interferometry has become a powerful means of modern military and civilian remote sensing information acquisition.For the single-baseline polarimetric SAR interferometry(SB-Pol In SAR), the mismatch of scattering process model in applications limits its ability to obtain the surface information of Earth surface. The practical requirements of SB-Pol In SAR emerge into the thorough analysis and comprehension of the relationship between system parameters, scene parameters and the observations, as well as the excellent methods of signal processing for the improvement of parameters inversion. With the massive observations from radar remote sensing, the addtion of space channels in the across-track direction enriches the information obtained by the polarimetric SAR interferometry. As an important development of modern imaging radar, three dimensional(3-D) reconstruction can be effectively achieved by the multi-baseline polarimetric SAR interferometry and tomography, as well as the vertical structure detection of complex natural scenes. For the complex scattering of targets, however, the dependence of observations on targets’ scattering characteristics is difficult to be determined. In the 3-D reconstruction, tomography is mainly subject to the baseline optimization and its implementation, as well as the error influence casued by the traditional two-dimensional SAR processing. Therefore, a robust and efficient focusing technique with high precision is an imperative for the enhancement of tomographic focusing performance.Aiming at these problems, this thesis investigates the robustness and precision enhancement of the signal and information processing in polarimetric SAR interferometry and tomography. The main research contents in this paper consist of fourkey points, that is, the scattering model mismatch investigation, the forest parameters inversion with SB-Pol In SAR, the construction of linear variation model(LVM) distribution model for double-baseline(DB) observations and volume scattering ambiguity, the baseline optimization and technique investigation for tomography. Supported by the xxx pre-research projects "Investigation of xxx system for information acquisition", "xxx technology with xxx satellites" and the xxx detection technology project "xxx technologies and experimental research", this thesis studies the forest parameters inversion with DB/SB-Pol In SAR and tomography, mainly around the scattering model mismatch, forest parameters inversion, volume scattering ambiguity-resolving and the baseline optimization for tomography. The main contents can be summarized as the following four parts:The first part investigates the scattering model based on the system and scenario characteristic parameters. At first, the mismatch of the scattering model is introduced before the analysis of the dependence of observations on system parameters and scenario features, including vegetation and terrain parameters. For this, the impact of terrain slope on the forest parameters inversion is determined according to the scattering model. As a result, the mapping is established for the description of the relationship between observations and the corresponding system/scenario parameters. in the model level, completely solve the mismatch problem caused by terrain model.The second part focuses on the forest parameters inversion with SB-Pol In SAR. Firstly, the method of the interferometric phase estimation is proposed for the underlying ground topography. According to the estimated topographic phase, the thesis presents a data-driven nonstationarity compensation for amplitude and phase auto-calibration. In view of the non-uniform distribution of observations, a technique called mapping space regularization is proposed. In addition, we investigate the invalid reference determination with a single-channel observation as prior information. To this end, the maximization of ground-to-volume ratio difference is first presented for the prior information fusion. Finally, the principal component analysis method is introduced to improve the accuracy and robustness of linear variation model.The third part of this thesis studies on the forest parameters inversion with dual-baseline polarimetric SAR interferometry(DB-Pol In SAR). The distribution model construction of the observations is investigated for DB-Pol In SAR. On this basis, a volume scattering ambiguity-resolving technique is proposed and validated through the analytical simulation. Then, a clustering phenomenon of observations is investigated for DB-Pol In SAR. Finally, a consistency-criterion-based algorithm for forest parameters inversion is proposed. To validate the effectiveness of the proposal, some experiments are carried out with different length ratio of baselines. The performance and error tolerance analysis shows that this method can effectively resolve the volume scattering ambiguity.The fourth part is dedicated to the baseline optimization and the 3-D tomographic imaging. For the purpose of non-uniform baseline optimization, the sweeping invariance of the array near field pattern is analyzed firstly to construct the optimization model for airborne SAR tomography. To solve the model, an objective function rasterisation is proposed and then the original model can be approximated as a Minmax optimization problem, which can be easily figured out. At last, an optimal-baseline-based tomographic imaging framework is presented for the natural scenario 3-D reconstruction. Based on the orthophoto correction technique, the 3-D representation of the tomographic images is implemented under the geodetic coordinates with the unitary transformation of coherency matrix.