ESA’s BIOMASS mission, successfully launched in April 2025, operates a first space born P-band Synthetic Aperture Radar (SAR) to measure global Above Ground Biomass (AGB) and retrieve Digital Terrain Models (DTMs) beneath dense canopies. Reliable interferometric terrain measurements critically depend on mitigating tropospheric prop agation delay. This thesis develops a robust and lightweight two-stage correction that is designed to avoid reliance on external datasets. First, the elevation-correlated stratified component is estimated directly from the interferometric phase and a Digital Elevation Model (DEM). Second, the residual turbulent component is reconstructed by interpolating phase values from high-coherence calibration points in open areas such as forest clearings or bare-ground patches, testing several interpolation strategies under sparse sampling. The algorithm is evaluated on L-band SAOCOM-1B data acquired over two contrasting environments: the arid, high-coherence Emi Koussi volcano in Chad, and the humid, vegetated Mount Fuji in Japan. Its performance is benchmarked against the state-of the-art Generic Atmospheric Correction Online Service (GACOS). The central question addressed is under which conditions a scene-specific DEM-based correction can rival or complement GACOS in practice, thereby informing atmospheric mitigation approaches for long-wavelength missions such as BIOMASS.
La missione BIOMASS dell’ESA, lanciata con successo nell’aprile 2025, è il primo SAR spaziale operante in banda P progettato per misurare la biomassa globale fuori terra (Above Ground Biomass) e per ricostruire Modelli Digitali del Terreno (DTM) sotto fitte coperture forestali. Misure interferometriche affidabili del terreno dipendono in modo critico dalla correzione dei ritardi troposferici di propagazione. In questa tesi viene sviluppata una pipeline di correzione robusta e leggera, concepita per evitare la dipen denza da dati ausiliari esterni. Nella prima fase, la componente stratificata, correlata con l’elevazione, viene stimata direttamente dalla fase interferometrica e da un Modello Digitale di Elevazione (DEM). Nella seconda fase, la componente turbolenta residua viene ricostruita interpolando i valori di fase da pixel di calibrazione ad alta coerenza in aree aperte, come radure forestali o zone di suolo nudo, testando diverse strategie di interpo lazione in condizioni di campionamento sparso. L’algoritmo è stato valutato su dati in banda L del satellite SAOCOM-1B acquisiti in due ambienti fortemente contrastanti: il vulcano Emi Koussi in Ciad, caratterizzato da condizioni aride e alta coerenza, e il Monte Fuji in Giappone, caratterizzato da forte umidità, vegetazione densa e turbolenza atmos ferica. Le sue prestazioni sono state confrontate con il servizio di correzione GACOS, considerato lo stato dell’arte. La domanda scientifica centrale affrontata è in quali con dizioni una correzione basata sul DEM e specifica per la scena possa eguagliare o integrare GACOSnella pratica, fornendo indicazioni per le strategie di mitigazione atmosferica nelle missioni a lunghezza d’onda lunga come BIOMASS.
Tropospheric correction strategies for low frequency SAR interferometry
Antonova, Elizaveta
2024/2025
Abstract
ESA’s BIOMASS mission, successfully launched in April 2025, operates a first space born P-band Synthetic Aperture Radar (SAR) to measure global Above Ground Biomass (AGB) and retrieve Digital Terrain Models (DTMs) beneath dense canopies. Reliable interferometric terrain measurements critically depend on mitigating tropospheric prop agation delay. This thesis develops a robust and lightweight two-stage correction that is designed to avoid reliance on external datasets. First, the elevation-correlated stratified component is estimated directly from the interferometric phase and a Digital Elevation Model (DEM). Second, the residual turbulent component is reconstructed by interpolating phase values from high-coherence calibration points in open areas such as forest clearings or bare-ground patches, testing several interpolation strategies under sparse sampling. The algorithm is evaluated on L-band SAOCOM-1B data acquired over two contrasting environments: the arid, high-coherence Emi Koussi volcano in Chad, and the humid, vegetated Mount Fuji in Japan. Its performance is benchmarked against the state-of the-art Generic Atmospheric Correction Online Service (GACOS). The central question addressed is under which conditions a scene-specific DEM-based correction can rival or complement GACOS in practice, thereby informing atmospheric mitigation approaches for long-wavelength missions such as BIOMASS.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_10_Antonova_Thesis_01.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Thesis text
Dimensione
47.75 MB
Formato
Adobe PDF
|
47.75 MB | Adobe PDF | Visualizza/Apri |
|
2025_10_Antonova_Executive_Summary_02.pdf
accessibile in internet solo dagli utenti autorizzati
Descrizione: Executive Summary
Dimensione
3.52 MB
Formato
Adobe PDF
|
3.52 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/243502