22 research outputs found
On the Phase Calibration by Multisquint Analysis in TOPSAR and Stripmap Interferometry
No full textThe availability of accurate trajectory information is paramount for the processing and exploitation of synthetic aperture radar (SAR) data. Considering the particular case of spaceborne SARs designed for repeat-pass interferometric applications, errors in the trajectory translate into phase artifacts that affect the interferometric performance. In this paper, we propose a model-based procedure to calibrate the trajectories of spaceborne SAR systems by the multisquint (MS) phase. The technique allows to estimate the along and the derivative of across track geometric errors. The geometric model of the InSAR phase is derived as a function of positioning errors and the MS phase model as derivative of the InSAR phase geometric model, with respect to the squint angle. We perform a sensitivity analysis of the model in order to define which geometric errors can be estimated by the MS phase, justifying the assumption that the MS phase is very poorly affected by the atmospheric phase screen. We particularly concentrate on the TOPSAR acquisition mode, where the phase is very sensitive to geometric errors. We start from the classical two-image case and then consider the extension to the multibaseline case. Experimental results obtained by processing of interferometric pairs acquired by the Sentinel-1A sensor are reported
Multi-Squint Analysis to Separate Geometric and Atmospheric Phase Artifacts in Spaceborne InSAR
No full textThe main issue during phase calibration of spaceborne
SAR system is to properly identify and separate
different contributions to the interferometric phase. In
this paper we suggest to exploit the Multi-Squint (MS)
interferometric phase in order to remove InSAR fringes
due to a linear orbital error, under the key assumption
that the MS phase is very poorly affected by
contributions from the atmospheric delay. Preliminary
results obtained by processing TerraSAR-X stripmap
data appear to confirm the validity of this assumption,
and suggest that MS processing can be operationally
employed for the calibration of spaceborne InSAR datastacks
A novel method for noise equivalent sigma nought estimation
No full textIn Synthetic Aperture Radar, Noise Equivalent Sigma Nought (NESZ) is usually verified by measurements over targets with ideally null backscatter, like calm water, or flat surfaces. This condition is quite different from the normal one, in term of quantization, source temperature, ambiguities and received signal energy. We propose a novel method based on interferometry that estimates NESZ in presence of a significant source backscatter. It exploits the principle that noise is the ultimate and unavoidable factor limiting coherence. For each value of backscatter (say in an histogram), a search is made for the target with highest coherence. NESZ is then estimated by fitting the theoretical model of coherence versus signal-to-noise-ratio. The estimate is performed for different range by exploiting narrow strips. The method relies on the existence of some target whose coherence is limited only by thermal noise. Results of processing Sentinel 1A / 1B data over the'Salar de Uyuni' site are here compared with the classical measures over signal-free region
«Experimental assessment of the PS-cal technique over COSMO-SKYMED high resolution SAR data
No full textThe purpose of the paper is to estimate the Azimuth Antenna
Pattern by using Persistent Point Scatterers (PPS) selected in
a stack of interferometric Synthetic Aperture Radar images.
Wide angle antenna pattern is estimated from targets spectra
by means of a Maximum Likelihood formulation. PPS are
considered a restricted subset of the Persistent Scatterers, for
which the point-shape feature is further required.
A model for PPS is provided and experimentally validated;
antenna estimation results are presented for both simulated
and real X-band Cosmo Skymed data
Geometric parameters retrieval in SAR systems
No full textL’obbiettivo di questa ricerca è analizzare gli effetti degli errori geometrici sui dati SAR, ed in particolare sulle applicazioni SAR che sfruttano due o più acquisizioni della stessa scena. Per far questo la tesi fornisce un modello teorico che mette in relazione gli errori di posizionamento relativi tra due acquisizioni e gli errori della geometria nominale. La formulazione matematica può essere applicata sia ad acquisizioni interferometriche che radargrammetriche, dato che gli errori di posizionamento relativi possono essere misurati sia tramite la fase interferometrica sia come errori di registrazione tra immagini. L’analisi del modello teorico ha permesso di identificare le componenti degli errori di geometria che possono essere stimati nelle diverse configurazioni dei sistemi SAR e l’accuratezza raggiungibile. Le analisi riportate in questa tesi si occupano principalmente di sistemi SAR satellitari.
La tesi può essere fondamentalmente divisa in due parti: la prima parte è dedicata alla derivazione del modello teorico e alla sua verifica con dati simulati. Mentre nella seconda parte il modello è stato impiegato in due differenti applicazioni: una applicazione di calibrazione interferometrica e un’applicazione di raffinamento del modello digitale del terreno tramite radargrammetria. I risultati ottenuti con tre diversi sistemi satellitari hanno dimostrato la validità del modello e la sua applicabilità.The objective of this research is the analysis of the effects of the geometrical error on the SAR data, focusing on the SAR applications that consider the differences between SAR acquisitions of the same scene.
In order to do this the dissertation provides a theoretical model by defining the relationship between the difference of the target positioning in two SAR acquisitions and the error of the nominal with respect the actual geometry.
The mathematical formulation is applicable both to interferometric and non-interferometric acquisitions; considering that the positioning difference can be measured either as Interferometric (InSAR) phase or as mis-registration shift according to the applications.
The analysis of the theoretical model allows to evaluate which components of the geometrical error can be estimated for different configurations of the SAR system, and the theoretical accuracy achievable.
All the analyses included in this thesis mainly deal with spaceborne SAR applications.
The first part of the dissertation is devoted to the derivation of the theoretical model of the geometrical errors and its verification with simulated data.
In particular, this part defines a mathematical formulation of the effects of the geometrical errors on the mis-registration shift in Line-Of-Sight (LOS) and Along-Track (AT) directions in case of interferometric acquisitions.
Concerning the non-interferometric case, the LOS mis-registration only will be dealt with.
At the end of this part, the InSAR phase contributions due to the Atmospheric Phase Screen (APS) and the target displacement will be analyzed in order to investigate their impact on the geometric parameters retrieval.
Two different SAR applications that exploit the suggested theoretical model will be described in the second part of the thesis.
The first application uses the AT mis-registration, in the field of the interferometric calibration, to validate the accuracy of the orbit products.
As a case study, the results of the orbit error retrieval will be shown for the Sentinel-1 satellite with a particular attention to the TopSAR acquisition mode.
Finally, the second application exploits the theoretical model in order to estimate the Digital Elevation Model (DEM) error by using non-interferometric acquisitions.
In particular, the reported case study refines the SRTM DEM by an iterative radargrammetric technique applied to high resolution SAR images in X-band.
The radargrammetric DEM generated by the suggested technique will be then compared to a photogrammetric DEM available for the area.DIPARTIMENTO DI ELETTRONICA, INFORMAZIONE E BIOINGEGNERIATelecommunications28GENTILI, GIAN GUIDOBONARINI, ANDRE
Tecniche di super risoluzione per tomografia SAR 3D in ambiente urbano
No full textLAUREA SPECIALISTICATra i principali orizzonti di ricerca nell’ambito del SAR sicuramente troviamo la tomografia SAR 3D, che tenta di risolvere i target in tutte e tre le direzioni spaziali. I dati a disposizione per svolgere la tomografia 3D sono solitamente immagini SAR 2D già focalizzate, che illuminano una stessa area da angolazioni differenti, in quanto sono ottenute da differenti acquisizioni dello stesso sistema SAR. Questi sistemi sono definiti Multipass e presentano particolari problematiche dovute fondamentalmente al numero limitato dei passaggi a disposizione e alla spaziatura irregolare in direzione di elevazione. Questa tesi analizza queste problematiche e propone delle tecniche che permettano di attenuarne gli effetti negativi che ne derivano sulle immagini tomografiche. Per far questo sono state svolte analisi di prestazioni su dati sintetizzati e successivamente generate sezioni tomografiche della scena a partire da immagini SAR già focalizzate di dati reali sia di un ambiente boschivo che urbano. Le tecniche proposte essendo pensate per focalizzare i target tipici dell’ambiente urbano hanno dato buoni risultati su questi dati. Come era intuibile le tecniche proposte non sono adatte a risolvere target tipici dell’ambiente boschivo
