Large-Scale SAR Image Simulation using EMPRISE® Software Compared to Real Image acquisitions with SETHI airborne Platform Over the Sainte-Marie-la-Mer countryside
Year: 2024
Type: conference-paper
Abstract: This study presents a novel approach to Synthetic Aperture Radar (SAR) image simulation and calibration, leveraging the EMPRISE® software against actual aerial data acquired by the SETHI airborne platform over the Sainte-Marie-la-Mer countryside. Aimed at bridging the fidelity gap between simulated and real SAR images, we introduce a localized quantile transformation technique that aligns the distribution characteristics of simulated imagery with those captured by SETHI. This process is validated through the Cumulative Distribution Function (CDF) distance metric, ensuring high accuracy of simulations in reflecting real-world data. Key to our approach is the ability not only to enhance the simulation of known areas but also to extend predictive capabilities to regions outside the SETHI capture scope. Results from applying our calibrated model to new areas confirm the efficacy of our methodology, showcasing a significant improvement in simulation fidelity and predictive accuracy. This work underscores the potential of advanced simulation techniques in remote sensing, offering a robust tool for enhancing SAR image simulations and extending their applicability to uncharted territories.
Authors: Thomas Houret, Olivier Lévêque, Benoit Talibart, Stéphane Roussel, Etienne Everaere, Kevin Unger, Nicolas Trouvé
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4404531672