Science Service System

Summary of Proposal CAL1744

TitleTSX radiometric calibration, phase calibration and pointing stability assessment by natural targets
Investigator Giudici, Davide - Aresys srl, SAR R&D area
Team Member
Prof. Monti Guarnieri, Andrea - Politecnico di Milano, Department of electronics and information DEI
Eng. Mancon, Simone - Politecnico di Milano, Department of electronics and information DEI
SummaryThe activity described within the current proposal would fall into the Calibration and Validation category foreseen by the AO. The analysis would allow the testing of the long term calibration and stability analysis technique (PS-CAL) on TSX data. The expected result is a set of performances achievable in the real case, but also the development of an efficient algorithm. 1.1.1Objectives ·Assess the capabilities of radiometric calibration, phase calibration and estimation of antenna pointing of the coherent and non coherent PS-CAL approaches with TSX SAR through exploitation of the stable pixels identified in a stack of SAR data. ·Study,implement and evaluate techniques for phase calibration of interferometricstacks. The scientific validation to be performed through TSX SAR data through a stack of SAR data The expected outputs are: The PS-CAL results on the stack: Radiometric normalization: Relative amplitude calibration for each image (amplitude normalization in azimuth and slant range). The temporal evolution of the estimated normalization constant. pointing stability: the estimated off-nadir correction (one value for each image). Orbital refinement: Refined PS elevation with respect to the ellipsoid. Baseline estimated with centimetric accuracy. Refined orbital state vector. Estimated phase field components and phase stability: atmospheric phase screen. Phase due to orbital errors. Phase due to the processing errors. and phase quality measure represented by quality maps. Funding: This study would be funded as a research project within the DEI - POLIMI.
Detailed reportSAR images as outputted by a typical SAR software, say single look complex format, are the result of the scene observed by the RADAR, the acquisition by the SAR instrument and, processing. The output is then used for a wide variety of applications that exploit amplitude, phases or both. The interest for the final users is exclusively in retrieving parameters closely related to the geophysical property of the scene: all those elements that represent the added value of SAR. In that the users dealing with amplitudes wish perfectly calibrated data, whereas users dealing with stacks and interferometry wish accurately coregistered data with all the predictable information removed, like fringes due to baseline errors. Amplitude and phase calibration have been studied since the early years of SAR. For each single image, phase measures the sensor-target optical path that is the optical distance between the phase centers of the sensor and the target. In a SAR image, phase is influenced and, consequently, carry information about: 1. the precise sensor orbit, 2. the local DEM 3. the atmospheric phase screen 4. local subsidence 5. volumetric effects (like vegetation or moisture) processing artifacts, likes coregistration errors The user is interested mainly in (4) and (5), and his wish is having all the other systematic elements removed, and the errors evaluated. On the other hand, users dealing with amplitudes need data to be precisely calibrated, both absolutely and relatively. To get the goal of amplitude and phase calibration, system and instrument designer implement several strategies: - internal calibration: related to test signals that are automatically generated by the instrument itself and routed into the system circuitry at regular intervals. The example for Sentinel-1 is represented by the RFC-mode pulses generated within a dataset acquisition - external calibration: related to measurements of external, known references. Usually most of the work is performed through internal calibration, aiming at ensuring the continuity of both radiometric and phase calibration through all the beams, and the acquisition time. Then external calibrators are used for fixing the absolute amplitude, whereas absolute phase calibration is not usually performed in the case of space-borne SAR, due to the uncertainty of the phase and the huge travel path contribution with respect to the wavelength. Geometric location accuracy is usually checked with external calibrators and consistency measures carried out by repeated pass images. The known targets used for external calibration are typically represented by corner reflectors (trihedral) or by active instruments (transponders) and by homogeneous targets (rain forest). Conversely, in this document we report results of external calibration and accuracy analysis (for both amplitude and InSAR phase) without any known targets. The PS-Cal aims to monitor the long term radiometric stability of a SAR system from the commissioning phase throughout the whole mission life and using independent targets from almost all-around the world. In this document we account for the proper elevation antenna pattern that is assumed known - say from internal calibration and antenna model - but for a pointing error, that may come from attitude uncertainty. As well the MS analysis exploits natural targets that remain coherent almost in whole azimuth bandwidth. We selected the MS analysis according to the key assumption that APS and DEM error contributions can be neglected from MS phase. The cumulonimbus is the kind of cloud which mainly affects the SAR signal, Highest level of Liquid Water Content (LWC > 0.5 g/m3). Even if top of the cloud is ~12 Km of altitude, the separation between two trajectories is 55 m. We compared the MS phase with the unwrapped InSAR phase clearly affected by APS. The MS phase is less affected by phase artifacts due to the APS. First we perform the SVD analysis of the forward model. The first two singular values go in the direction of the following three parameters: Mis-Registration in x direction, Orbit error variation in y direction, Orbit error variation in z direction. After the application of the PS-CAL and the MS phase methods to the provided TSX dataset, we can conclude that: The MS phase has been demonstrated as a valuable tool for InSAR phase calibration: It is minimally affected by APS and DEM It senses orbit errors The MS phase senses along-track displacements Analyses in TSX STRIPMAP and RADARSAT-2 TOPSAR geometries have shown that MS phase can be described by a model with two singular values. The model is capable to detect bursts affected by low coherence or local along-track displacement. TOPSAR artifacts due to orbital errors are so removed. Artifacts can be removed as well by adaptive estimation and compensation of TOPSAR phase jumps, but such technique would not be able to cope for decorrelation and local deformations. The extended PS-Cal results confirm the high radiometric stability of the TerraSAR-X data in terms of roll-angle and image amplitude (a linear trend of less than 0.07 dB/year was estimated). The measures of orbit accuracy confirm the specifications of the science orbit products with a measured 1-sigma 3D accuracy of 20 cm.

Back to list of proposals

© DLR 2004-2016