Development of Ionosphere Estimation Techniques for the Correction of SAR Data

Abstract

Low frequency synthetic aperture radar (SAR) configurations play an important role in monitoring forests and ice sheets on global scales. They provide the ability to image systematic large areas at high spatial and temporal resolution, while at the same time they allow to penetrate into the forest and ice volume providing information about the vertical structure of such media. However, at lower frequencies spaceborne SAR measurements are affected by the presence of the ionosphere, i.e., a dispersive, anisotropic and spatial as well as temporal inhomogeneous partial plasma, located in the upper part of Earth's atmosphere. The ionospheric layer is characterised by an anisotropic (spatial and/or temporal) variation of the refractive index that affects intensity, polarisation as well as (phase and group) velocity of the transmitted and received SAR pulses passing through. In consequence, the amplitude and phase of the finally obtained (complex) SAR images are distorted impacting intensity, polarimetric and interferometric SAR measurements. The ionosphere can be characterised in terms of the Total Electron Content (TEC), which stands for the integrated number density of the free electrons along the propagation path. A consistent theoretical framework, required for describing the ionospheric impact (by means of TEC) on intensity, polarimetric and interferometric SAR observables, has been established in this thesis that allows to reveal the interconnection between the ionospheric impact on the individual elements of different SAR observation spaces. A variety of techniques have been developed and proposed to compensate the effects, induced by the ionosphere by estimating the distortion on a single SAR image parameter (e.g. focus, contrast, geometry, or polarimetric signature) and correcting the associated SAR observable. Common to all these techniques is the fact that only the ionospheric distortion on a (small) subset of the available observation space is considered, while the remainder of the redundant and synergetic information is ignored. As a result, the achieved performance depends strongly on the ionosperic conditions as well as on the SAR system parameters. In this thesis an innovative integrated ionospheric compensation approach, that maximizes the exploited information, has been developed. An accurate high (spatial) resolution TEC map is obtained by combining the distortion estimates of a larger set of conventional, polarimetric and/or interferometric image parameters that is then used to compensate the ionospheric impact with in-creased accuracy and robustness across the whole observation space. This leads to optimized correc-tion results for a wide range of ionospheric conditions (including scintillations) and SAR system parameters. The proposed methodology has been implemented and assessed by means of a large set of single image or interferometric quad-polarimetric ALOS-PALSAR data and simulated BIOMASS P-band data covering a wide range of ionospheric conditions. The achieved performance is critically analysed and discussed in terms of ionospheric conditions and system or data parameters and then projected on different spaceborne SAR scenarios.