elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Contact | Deutsch
Fontsize: [-] Text [+]

A Bayesian Method for Very High Resolution Multi Aspect Angle Radargrammetry

Goel, Kanika (2009) A Bayesian Method for Very High Resolution Multi Aspect Angle Radargrammetry. Master's, Technische Universität München.

Full text not available from this repository.

Abstract

InSAR has been used for mapping of urban areas and buildings from space in three dimensions. In recent years, InSAR’s capabilities have been considerably improved by advanced processing techniques such as Persistent Scatterer Interferometry and SAR Tomography. However, complicated scattering situations such as layover need to be better resolved in urban areas. This master thesis presents a new technique for providing the 3D scatterer location and the resolution cell configuration and consequently, more precise surface displacement of urban areas can be realized by combining this technique with Persistent Scatterer Interferometry and SAR Tomography. The implemented method is based on the principle of radargrammetry and uses Bayesian inference to estimate heights of point scatterers. Since the method is based only on amplitude of SAR images, it is insensitive to limitations caused by coherency due to temporal decorrelation, a small spread of incidence angles caused by a small orbital tube, the atmospheric phase screen and non-linear displacement. . For implementing this method, a new estimation framework has been developed, wherein, directed graphs are used to represent dependencies of random variables and probability density functions are modelled by particle representations. The principle of radargrammetry is converted to a simple Bayesian line fitting problem for estimating heights of individual point scatterers. To improve the technique, joint estimation of small neighbourhoods has been implemented. The implemented estimation framework allows a Bayesian model selection, which is used during joint estimation of a small neighbourhood to test for the hypotheses of a vertical wall or a horizontal surface. High Resolution Spotlight TerraSAR-X preprocessed interferometric data of the Berlin central station has been used as a processing example. The results obtained for Berlin central station show that this new technique is capable of unambiguous height estimation of point scatterers. The proposed technique can take benefit from future SAR systems with 600 MHz chirp bandwidth.

Document Type:Thesis (Master's)
Title:A Bayesian Method for Very High Resolution Multi Aspect Angle Radargrammetry
Authors:
AuthorsInstitution or Email of Authors
Goel, KanikaKanika.Goel@dlr.de
Date:30 September 2009
Journal or Publication Title:Master Thesis
Number of Pages:67
Status:Published
Keywords:SAR radargrammetry, Bayesian inference, particle filters, directed graphical models, model selection, Persistent Scatterer Interferometry, SAR Tomography, TerraSAR-X
Institution:Technische Universität München
Department:Lehrstuhl für Methodik der Fernerkundung
HGF - Research field:Aeronautics, Space and Transport (old)
HGF - Program:Space (old)
HGF - Program Themes:W - no assignement
DLR - Research area:Space
DLR - Program:W - no assignement
DLR - Research theme (Project):W -- no assignement (old)
Location: Oberpfaffenhofen
Institutes and Institutions:Remote Sensing Technology Institute > SAR Signal Processing
Deposited By: Nico Alexander Adam
Deposited On:25 Feb 2010 13:08
Last Modified:25 Feb 2010 13:08

Repository Staff Only: item control page

Browse
Search
Help & Contact
Informationen
electronic library is running on EPrints 3.3.12
Copyright © 2008-2012 German Aerospace Center (DLR). All rights reserved.