Skip to main content
- Speaker: Patrick Haughey (NC State)
Time/Location: March 26, 1:30-2:30pm, SAS 4201
Title: ISLR-Guided Swarm Gradient Descent for Transionospheric SAR Autofocus
Abstract: Orbital synthetic aperture radar (SAR) systems operating at low frequencies such as L- or P-band are sensitive to the dispersive effects of the ionosphere. This research aims to mitigate the distortions of SAR images due to the scintillation phase errors (SPE) induced by ionospheric turbulence. Transionospheric autofocus encompasses the techniques used to estimate and compensate for unknown phase errors induced by the ionosphere. Successful implementation restores image contrast and mitigates geometric warping in the reconstructed SAR imagery. Previously, we developed a novel variational autofocus algorithm that retrieves the parameters of the turbulence from SAR data by optimizing a contrast-enhancing cost function and subsequently uses those to correct the distortions. The increased effectiveness of this autofocus method has been corroborated statistically in the case of a single gradient-based optimizer. Yet the primary limitation of our method is due to the highly non-convex nature of the cost function. In this work, we show that autofocus efficacy can be further improved with the help of global optimization. Most importantly, we demonstrate that a key measure of image quality — the integrated sidelobe ratio (ISLR), a non-differentiable but physically meaningful radar imaging metric — can be effectively included into the swarm-based gradient algorithm to achieve a statistically significant increase in performance.
Find NC State websites, locations and people
TOP LINKS