Calibration of Sentinel-1 and TerraSAR-X time-series SAR data over the Queensland Corner Reflector Array, Australia
Medhavy Thankappan1, Ulrich Balss2, Matthew Garthwaite1, Christoph Gisinger3, Peter Meadows4, John Peter Merryman Boncori5, Adrian Schubert6
1Geoscience Australia, 2DLR, 3Technical University of Munich, 4BAE Systems, 5Technical University of Denmark, 6University of Zurich
Queensland Corner Reflector Array
40 corner reflectors spread over 100km x 100km
Details of individual CR sites available from the Point and Distributed Targets Database maintained by the CEOS WGCV SAR Subgroup at http://sarcv.ceos.org/targets/target_group/4/
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
Performance Testing of CRs
Signal-to-Clutter Ratio for corner reflectors derived from TerraSAR-X, RADARSAT-2, Sentinel-1A and ALOS-2 SAR imagery. Thumbnail images show the 1.5 m corner reflector response in intensity images from each SAR sensor at site 6.
# SAR images # reflectors imaged Mean SCR Std. Dev.
TerraSAR-X 11 11 49.57 3.19Sentinel-1A 3 38 32.00 2.85
RADARSAT-2 8 40 36.18 2.17ALOS-2 11 36 29.86 6.14
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
Sentinel-1A: Radiometric Performance
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
HH
Radiometric analysis of 10 HH images selected for the period from 4th June 2016 to 3rd September 2017.
Relative RCS for corner reflectors with normalisation for each acquisition.
The overall rRCS standard deviation without and with normalisation is 0.38dB and 0.32dB respectively. HH
Sentinel-1A: Radiometric Performance
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
HH
Mean and standard deviation rRCS (no normalisation) for images acquired since the start of 2016.
Good rRCS results are obtained including the three 2.5m CRs (#3, #5 & #14) for EW mode (although the STD is larger than for IW mode).
Relative RCS (dBm2) Number of Measurements
IW All Corner Reflectors -0.05±0.51 1019
IW1 All Corner Reflectors -0.06±0.51 604
IW2 All Corner Reflectors -0.05±0.52 415
IW 2.5m Corner Reflectors -0.18±0.22 77
EW 2.5m Corner Reflectors 0.21±0.67 40
Relative RCS (dBm2) Number of Measurements
IW All Corner Reflectors -0.17±0.40 126
IW1 All Corner Reflectors -0.19±0.37 73
IW2 All Corner Reflectors -0.14±0.44 53
IW 2.5m Corner Reflectors -0.20±0.07 77
VV
Sentinel-1B: Radiometric Performance
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
HH
Radiometric analysis of 10 HH images selected for the period from 8th October 2016 to 21st September 2017
Relative RCS is plotted for corner reflectors with normalisation for each acquisition.
The overall rRCS standard deviation without and with normalisation is 0.32dB and 0.28dB respectively.
HH
Sentinel-1B: Radiometric Performance
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
HH
Mean and standard deviation rRCS (no normalisation) for images acquired since the
start of the S1-B routine phase (22nd September 2016).
Good rRCS results are obtained including the three 2.5m CRs (#3, #5 & #14) for EW
mode (although the STD is larger than for IW mode). All HH polarisation.
Relative RCS (dBm2) Number of Measurements
IW All Corner Reflectors 0.09±0.42 620
IW1 All Corner Reflectors 0.14±0.40 371
IW2 All Corner Reflectors 0.01±0.44 249
IW 2.5m Corner Reflectors -0.13±0.18 47
EW 2.5m Corner Reflectors 0.38±0.74 20
Sentinel-1 ALE: Predicted – Observed Positions
Slant-range
µIW1 = -0.17 m
sIW1 = 0.30 m
µIW2 = -0.43 m
sIW2 = 1.67
µAll = -0.28 m
sAll = 1.14 m
Azimuth
µIW1 = 2.23 m
sIW1 = 0.36 m
µIW2 = 1.11 m
sIW2 = 1.02
µAll = 1.74 m
sAll = 0.92 m
Slant-range
µIW1 = -0.03 m
sIW1 = 0.21 m
µIW2 = -0.02 m
sIW2 = 0.31
µAll = -0.028 m
sAll = 0.26 m
Azimuth
µIW1 = 2.26 m
sIW1 = 0.32 m
µIW2 = 1.48 m
sIW2 = 0.24
µAll = 1.92 m
sAll = 0.48 m
Positioning results after updates (Nov.2016)Positioning results prior to updates
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
Three dates analyzed (09 Aug., 26 Sep., 08 Oct, 2015)
Sentinel-1: Absolute Location Error (ALE)Results prior to positioning update Results after update
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
Sentinel-1A IW (TOPS) products from 2014 10 30 to 2017 08 10 (56 passes in total)
TerraSAR-X / TanDEM-X: ALE
No mount model adjustment With mount model adjustment
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
TanDEM-X ScanSAR products from 2017 04 18 to 2017 09 08 - 7 passes in total
Sentinel-1 ALE: Intra-burst-dependent Range Bias
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
• Improvement of range ALE for TOPS products after implementation of a range correction1 dependent on the target intra-burst position
• Improvement of bistaticcorrection
1Rodriguez-Cassola M., Prats-IraolaP., De Zan F., Scheiber R., Reigber A., Geudtner D., Moreira A. Doppler-related distortions in tops SAR images. IEEE Trans. Geosci. Remote Sens. 2015, 53, 25–35.
Before After
Sentinel-1: Old vs New Bistatic Corrections IW / EW
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
Old OldNew New
Summary
An array of 40 CRs of 1.5, 2.0 and 2.5m sizes deployed in Queensland, Australia; some CRs individually characterised for RCS; designed to support SAR calibration
Australian CR array supporting calibration of the Sentinel-1 constellation under the ESA FRM4SAR initiative
Radiometric consistency of CRs evaluated for Sentinel-1A/B IW and EW modes, and found to be stable
Sentinel-1 geometric calibration : three independent assessments for S1 have shown improvement in ALE following updates to CR positions and other corrections
ALE for TerraSAR-X/TanDEM-X over CRs improved with use of mount model
CEOS SAR Calibration/Validation Workshop 2017, Pasadena, 7-9 November
Thank you