Title: A report to the 22nd CEOS CALVAL from CEOS SAR CAL VAL subgroup based on discussion at 12th workshop
1A report to the 22nd CEOS CAL/VAL from CEOS SAR
CAL VAL subgroup based on discussion at 12th
workshop in ULM Germany
Masanobu Shimada JAXA/EORC
2- SAR subgroup
- Mission to foster high-quality synthetic aperture
radar imagery from airborne and spaceborne
systems through precision calibration in
radiometry, phase, and geometry, and validation
of high level products - Objectives
- Act as a forum for international technical
interchange on the evolving methodologies,
techniques, and equipment of SAR calibration and
validation. - To determine standard definitions and
calibration-validation requirements for
synthetics aperture radar imaging systems. - To support changes in CEOS formats and user
products as appropriate. - To facilitate international cooperative programs
in the calibration and validation of SAR systems. - To educate the SAR community
- Action plans
- Annual meeting
- Set up standard cal/val sites -gt inter-sensor
comparison - Calibration accuracy requirement for Polarimetry
gt2003
DONE.
DONE.
on going
3Summary of the 12th CEOS SAR CAL/VAL in ULM
12th subgroup workshop, organized by ESA, has
been successfully held on May 27 and 28 2004 at
VDE/ULM/Germany. It collected more than 38 papers
(28 oral and 10 poster presentations), and 70
participants from various countries. They focused
on the topics of geometry and radiometry,
Oceans and atmosphere, Calibration
methodology, Polarimetry and Interferometry,In
strument Calibration Concepts, and Land. Each
session had 20 minutes discussion time for the
seed questions, which are prepared among the
participants. Session summary -gt each session
chairman make 12 pages and send Chris Buck
4- Geometry and radiometryERS-2,Radarsat,Geometric
Cal of Envisat-ASAR,Bistatic SAR, - Oceans and atmospherewave spectrum, wind
vector, Siberia-II - Calibration methodology ASAR calibration, ARC
use for ASAR cal, POLCAL of PALSAR, etc. - Polarimetry and Interferometry Land cover
detection and tree height estimation by
POL-IN- SAR, Calibration performance of
X-SAR/Australia - Instrument Calibration ConceptsDesign
performance of Advanced PARC, In-orbit
characterization of Radarsat-2, etc. - LandTerra-sar-L, etc.
5Image Quality Historyof CDPF Products
Focusing (PSLR) All Beams
Energy Concentration in main lobe (ISLR) All Beams
Increased solar activity
Geometrical Registration (ALE) All Beams
Radarsat-1 after COTE
6Radiometric Calibration
Radarsat-1 after COTE
- Monitoring
- Characterization of elevation gain pattern in
terms of radar reflectivity - Radiometric accuracy comparison of measurement
with calibration pattern
- Antenna Pattern Changes ("Radiometric Anomalies")
- Have occurred gradually since 1998
- Cause
- Experiment investigation on Variable Phase
Shifters (VPSs) in 2001 - Most likely due to changing performance over time
- not temperature - Corrective Action ?o recalibration of affected
beams - Comparison of new measurements with current
calibration pattern - Updating payload parameters file with
recalibrated data - Peak-to-Peak Deviation lt 2.2 dB in worst case for
whole swath - Design goal of 1.0 dB
7Radiometric Calibration
Radarsat-1 after COTE
Radiometric Accuracy Beam F5 Never recalibrated
Radiometric Accuracy Beam W1 Recalibrated twice
Elevation Beam Pattern for W1
8ERS-2 after Meadow
- An additional source of point targets for quality
assessment are ground stations. As these acquire
ERS SAR raw data in real time, they will be
pointing towards the satellite - the Kiruna and
Neustrelitz ground stations have been used
Kiruna (12.5km by 12.5km)
Neustrelitz (3km by 3km)
9UL -7.068, -72.100 UR -7.182, -70.265 LL
-10.668,-72.892 LR -10.786,-71.051
Sigma-naught stability of Amazon forest
PALSAR Calibration sites in Amazon
10- Recommendation
- 1) Space agencies to make best efforts to
generate high quality SAR products, considering
that the achievable geometric accuracy is the
order of 15m or less, and radiometric accuracy is
significantly less than 1.0 dB. - 2) Space agencies to consider to generate user
friendly SAR products, for example to add image
quality information(i.e., ambiguity, saturation,
shadow,,,) to the current SAR products. - 3) Space Agencies should continue to improve the
quality of single and full polarization SAR data
by technical exchange of information on the
development of SAR calibration and imaging
algorithms and techniques.
11- What we decided at the 12th CEOS SAR CAL/VAL are
- 1) A standard calibration site will be jointly
setup by space agencies at Amazon area, selecting
such an area that collocates from their existing
sites (and validation site of land product
group). The tentatively agreed area has
coordinates of (UL -5.03, -65.67 - LR -9.12,
-69.64) -gt See Fig.1 - 2) Space agencies to apply the site for
calibrating their SAR, i.e., antenna elevation
pattern, sensor stability monitoring, absolute
calibration, etc., and make an annual report at
the SAR CAL/VAL. This action almost responds to
the action plan of CEOS CAL/VAL, i.e., SAR data
should be traceable to the international
standard. - 3) Data formats are now not compatible among
satellites. We received some requirements on
keeping the format compatibility.
12JAXA UL -7.068, -72.100 UR -7.182, -70.265 LL
-10.668,-72.892 LR -10.786,-71.051
International SITE UL -5.03, -65.67 UR -5.03,
-69.64 LL -9.12, -65.67 LR -9.12, -69.64
ESA UL -4.000, -71.000 UR -4.000, -65.000 LL
-11.000,-71.000 LR -11.000,-65.000
CSA UL -6.06, -68.28 UR -6.06, -66.33 LL
-7.24, -68.28 LR -7.24, -66.33
EOS Land Validation (LBASantarem) Center Lat,
Lon -2.857, -54.959
13- 4) Geometric and radiometric accuracy
requirements highly depend on the user
requirements and retrieval algorithm. - 5) These requirements will be reviewed and
updated every year. - 6) Next workshop will be held in Australia.
14Action Item status
15- SAR subgroup to request WGISS to provide task
team for Pol-In-SAR data formatting. -gtIn
Toulouse workshop (1999), it was recommended that
SLC is the only data for preserving phase for
interferometry compressed data is necessary for
polarimetry. -gt In Montreal (2003), two
candidates were listed, SLC and the other
(preserving phase). -
- By communication within science team and
discussion at the workshop, it was decided that
Single Look Complex is the best adequate format.
162) The WGCV SAR subgroup should set up
calibration and validation reference sites (to
include both natural and man made targets) In
Montreal, we have confirmed the Amazons
stability for some test sites, 1 in Radarsat, 1
in JERS-1, and whole Amazon in ERS). We will list
the possible sites in 2004 meeting. The sites
location should be discussed in 2004. Since the
Amazon area is the stable site for C and L band
SAR. Land product group has some validation site
in Amazon (not connected with this site). -gt
The site is set at south west of Amazon.
173) Space Agencies should continue to improve the
quality of single and full polarization SAR data
by technical exchange of information on the
development of SAR calibration and imaging
algorithms and techniques. gtThis action is
being conducted through the several international
science workshops and CEOS SAR CAL/VAL workshop.
-gt recommendation
184) The stability of the Amazon rainforest should
be systematically studied as a function of
incidence angle, frequency and polarization using
existing data sets including SEASAT, ERS, JERS,
RADARSAT and TRMM and emerging data sets from
ENVISAT, TerraSAR and RADARSAT-2. Funding to
complete the study should be solicited at the
CEOS Plenary. -gt (Finished). One question was
raised, "how does Envisat/ASAR measure the
stability of the Amazon? Also, we should think
of uploading the Amazon stability dataset
(gamma-zero, sigma-zero, incidence angle,
polarization) at the CEOS SAR CAL/VAL website so
that any people can access. Papers for
Radarsat-1, JERS-1, and ERS-1 in the Amazon
session of Montreal are adequate. Adding the
Envisat is hopeful. Workshop report was made
available by CSA in Dec. 2003. But, the ENVISAT
ASAR and ERS-2/1 AMI has not been compared. This
will be done and presented at CEOS SAR CAL/VAL
13th in Australia. (Half finished and half not
finished).
195) Next meeting of the SAR subgroup should focus
on the problem of full polarimetric SAR
calibration, both establishing requirements and
techniques. System stability (short and long
term)? Cross-Talk lt -35 dB Channel amplitude
imbalance 0.2 dB (soil moisture) Channel-phase
2?- 5? Channel registration lt0.1 pixel
? Requirements should be expressed in term of
NESZ as a function of the frequency X -23
dB? C -25 dB? L -30 dB? P -35 dB? L-band
Faraday rotation correction requirement 2.5
degree (it is accepted that C-band Faraday
rotation of 2.5 is not significant)
This has not been approved. We will rebuild the
table through the mail and internet by ESA, CCRS,
JAXA, etc.
206) Space agencies should ensure that calibration
processes for SAR are traceable to the primary
standards. -gt Ongoing.
21- Future Plan
- Proceedings distribution of 12th workshop on
CD-ROM by end of June
- Paper publishment
- Proceedings of 11th CEOS SAR CAL/VAL issued Dec.
2003 by CSA.
22CEOS SAR Calibration and Validation Workshop2004
- Seed Questions for Discussion
23Seed Questions Geometry and Radiometry
- What geometric accuracy should be achievable
using only the information from recent satellites
(Doppler, attitude data, position data, star
tracker)? - What radiometric accuracy should be achievable
using only the information from recent
satellites?
24Seed QuestionsOceans and Atmosphere
- Bearing in mind the exciting new products
available from SAR over the ocean and in the
atmosphere -
- Is the calibration accuracy for current SARs
still adequate? Could the specification be
relaxed, should it be extended? - Are new level-2 geophysical validation standards
and methods required for the new level-2
products?
25Seed QuestionsCalibration Methodology
- What measures can be taken to further streamline
the process of generating calibrated imagery? - Are steps being taken to provide information
relating to the level of ambiguous returns in the
imagery? Should this take the form of a 2-D
lookup table for ScanSAR products?
26Seed QuestionsPolarimetry and Interferometry
- What is the impact of the quality of polarimetric
calibration on classification and parameter
retrieval algorithms? - What is the impact of Faraday rotation on
polarimetry and POLinSAR applications? - In POLinSAR height estimation, what are the
calibration requirements for the two prime
algorithm approaches, namely DEM differencing and
coherence model inversion? - What is the status of calibration in bi-static
polarimetry? How do we set up problems to
validate the extra information in the HV/VH
difference channel?
27Seed QuestionsInstrument Calibration Concepts
- Advanced SAR concepts based on TRMs have a more
sophisticated on-board calibration scheme
providing individual TRM characterization as well
as in-orbit antenna characterization. What will
be the requirements for future on-board
calibration concepts? What are the advantages of
a calibration of the SAR data on board? - Rainforest regions and calibration transponders
are being used and are also planned to be used
for future SAR missions. New transponder concepts
offer coding of the pulses, quad-pol mode and
even an interferometric capability. How might
both types of calibration method (transponder and
rainforest) be combined in the most effective way
(i.e. reducing the total effort of the
calibration)? - Yaw steering has been used in most spaceborne SAR
missions for reducing the effective Doppler
centroid in the processing. What are the
limitations for not using a total zero Doppler
approach in future spaceborne SAR systems? - New algorithms are being developed for bi-static
SAR processing. What are the challenges for
bi-static SAR calibration?
285) Next meeting of the SAR subgroup should focus
on the problem of full polarimetric SAR
calibration, both establishing requirements and
techniques. gt Engineering requirement has been
almost (not fully) established. The scientific
requirement and techniques will be discussed in
future. The current requirements System
stability (short and long term)? Cross-Talk lt
-35 dB Channel amplitude imbalance 0.2 dB (soil
moisture) Channel-phase 2?- 5? Channel
registration lt0.1 pixel ? Requirements should be
expressed in term of NESZ as a function of the
frequency X -23 dB? C -25 dB? L -30
dB? P -35 dB? L-band Faraday rotation
correction requirement 2.5 degree (it is
accepted that C-band Faraday rotation of 2.5 is
not significant)
29Seed QuestionsLand
- To what extent are the applications presented
dependent on the choice of frequency and/or
wavelength? - To what extent are the applications presented
dependent on calibrated data?