Wilkinson, Andrew John;
(1997)
Techniques for 3-D surface reconstruction using synthetic aperture radar interferometry.
Doctoral thesis (Ph.D.), University College London (United Kingdom).
Preview |
Text
Techniques_for_3-D_surface_rec.pdf Download (36MB) | Preview |
Abstract
SAR Interferometry (InSAR) has emerged in recent years as a promising technology for topographic mapping. The classical approach to InSAR topographic reconstruction involves the decomposition of the problem into a number of distinct stages: SAR focusing, image registration, interferogram formation, phase unwrapping, baseline parameter estimation and phase-to-height conversion. The initial objectives of this research were to investigate and develop techniques for reconstructing topographic maps based on the classical decomposition, focusing particularly on the phase unwrapping and phase-to-height conversion problems. The thesis begins with a detailed review of existing InSAR theory and classical InSAR processing. This is followed by an analysis of the problem of reconstructing a height map from an unwrapped phase map. A generalized vector analysis of the InSAR geometry is developed and models for the sensor flight trajectories are proposed. Techniques from classical estimation theory are applied to obtain estimates of model parameters from calibration point data. Error propagation theory is applied to estimate the covariance matrix of reconstructed elevation points in terms of the covariance matrix of the calibration points and an estimate of the accuracy of the unwrapped phase. In the area of phase unwrapping, existing strategies are reviewed and analysed. From this investigation, it is concluded that currently employed algorithms are not model-based, and may be viewed as "ad hoc" classical estimators. Additionally, currently employed algorithms provide no measure of uncertainty in the reconstructed phase and consequently existing InSAR processors do not provide realistic indicators of the height uncertainty. This led to the development of a novel, probabilistic treatment of phase unwrapping as a model-based Bayesian inference problem. Consequently a new approach to InSAR topographic mapping is proposed, avoiding many of the conceptual difficulties encountered in the classical understanding of SAR interferometry. The approach is based on the treatment of the InSAR reconstruction problem as a model-based Bayesian inference problem, formulated in terms of a direct parameterization of the topographic surface. Models are proposed for the image formation process and for the surface elevation and reflectivity. The issue of formulating a measure of height uncertainty is addressed.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D. |
| Title: | Techniques for 3-D surface reconstruction using synthetic aperture radar interferometry |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Thesis digitised by ProQuest. |
| Keywords: | (UMI)AAIU643015; Applied sciences; SAR Interferometry |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10100302 |
Loading...
Loading...Loading...Loading...Archive Staff Only
 |
View Item |
