||Study on Multiple Angle Polarimetric Radar Image
Watanabe, Takuma渡邉, 卓磨
132 , 2015-03-23 , 新潟大学
学位の種類: 博士（工学）. 報告番号: 甲第4046号. 学位記番号: 新大院博（工）甲第428号. 学位授与年月日: 平成27年3月23日
This thesis deals with the analysis of multiple angle polarimetric synthetic aperture radar (PolSAR) images in the sense that the radar acquires the data from multiple incidence angles as well as azimuthal angles. The radar images are known to have strong dependence on these angular parameters and target physical or geometric characteristics such as the dielectric properties, roughness, shape, orientation, and so forth. Therefore, the multiple angle observation and analysis are necessary to aid the understanding of the PolSAR images. We carry out multiple angle PolSAR observation under well-controlled conditions in an anechoic chamber, and several characteristics of the multiple angle PolSAR images are revealed. The major contributions of this thesis are summarized as follows: 1. Experimental investigation on how soil moisture affects the polarimetric backscattering from forested terrain using the multiple angle measurement data. 2. Comparison of model-based polarimetric decomposition algorithms applied for PolSAR and tomographic SAR (TomoSAR) images. 3. Introduction of polarimetric concepts to circular synthetic aperture radar (CSAR) observation and image reconstruction. In the rst contribution, we examine the effects of soil moisture on the double-bounce scattering occurred between the ground surface and trunks. We conduct simple model experiments in an anechoic chamber, and show that the co-polarization ratio and phase difference can be used to estimate the moisture condition of forested terrain. The second contribution is the comparison of model-based polarimetric decomposition algorithms which are used to decompose a mixture of several scattering mechanisms into some simple scattering components. Although a number of decomposition algorithms have been developed, the physical interpretation of the decomposed results are still under investigation. To help the understanding of the decomposed results, we provide the qualitative comparison of these algorithms using the anechoic chamber measurement data. Also the introduction to the polarimetric decomposition schemes to TomoSAR data reveals the possibility of recovering the three-dimensional structure and scattering mechanisms of the observing targets. The last contributions are related to CSAR image analysis. Apart from the conventional SAR systems which have a linear flight path, CSAR has a circular trajectory to acquire the Scattering of the observing scene for high-resolution radar image reconstruction. The introduction of polarimetric concepts to CSAR data becomes challenging due to the diversity of the polarization basis. In this thesis, we deal with this problem by introducing polarization basis transformation to CSAR reconstruction algorithm. It is shown that the building structure can be recovered by this novel CSAR image reconstruction processing.