Loading...
Mitigation of artifacts in radar imaging using microlocal analysis with applications to passive imaging and environmental scattering
Citations
Altmetric:
Date
2025-12
Abstract
In the analysis of many synthetic aperture radar (SAR) experiments, an idealised model is considered that neglects the possibility for background signals to be recorded simultaneously and corrupt the image. This thesis addresses two scenarios where this can occur. In the first scenario we consider, there is “crosstalk” between the signal of the controlled emitter with that of an uncontrolled emitter (such as a radio tower, for example) present in proximity to the experiment. To investigate this phenomenon we analyse a multistatic SAR experiment where the measured data cannot be separated into contributions from individual emitters prior to image formation. The second scenario of consideration here, occurs due to multiply scattered waves. By incorporating a known environmental scatterer (such as the wall of a large building or a cliff face) into the model we can account for the different paths along which a backscattered wave can return to the receiver, which would normally be omitted in a more idealised analysis.
For each of the experiments considered in this thesis, the model for the radar data is given by a Fourier integral operator (FIO). FIOs can be analysed using tools from microlocal analysis. Using such methods, we show that in all of the cases in which background signals are present, there will be artifacts in the image, and we determine their locations relative to the scatterers that produced the data. To combat this, we develop methods that allow us to create an image of a region of interest (ROI), that is free from such artifacts. The first method makes use of a carefully designed data acquisition geometry to localise artifacts away from a region of interest and the second is an image processing technique that displaces artifacts away from an ROI.
Following this, we consider the scenario where the receiver can only make a single pass over the ROI. The result of this restriction is that the scattering operator is now a singular FIO. This presents difficulties when attempting to mitigate against artifacts in an ROI as we did previously. We discuss these limitations in detail and discuss some potential means by which a faithful reconstruction can still be obtained.
Finally, we use the methods of microlocal analysis that have been called upon throughout this thesis to investigate two more problems arising in SAR imaging. The first involves forming an image from bistatic radar data as if it had been generated by a monostatic transceiver. In the second, we incorporate a random variable into the model to account for phenomena such as turbulence during the experiment, for example.
Supervisor
Description
Publisher
University of Limerick
Citation
Collections
Files
ULRR Identifiers
Funding code
Funding Information
Sustainable Development Goals
External Link
License
Attribution-NonCommercial-ShareAlike 4.0 International