RADARSAT1 Image Quality Evolution to the Extended Mission

1
RADARSAT-1 Image Quality Evolution to the
Extended Mission
S. Cote, S. K. Srivastava, P. Le Dantec, R.
GrayCanadian Space Agency, Saint-Hubert, Canada
R. K. Hawkins, K. P. MurnaghanCanada Centre for
Remote Sensing, Ottawa, Canada
Stephane CoteManager, Data QualitySatellite
OperationsCanadian Space AgencyJohn H. Chapman
Space Centre6767 route de l'AéroportSaint-Hubert
, QC, Canada J3Y 8Y9Tel. [phone] Fax
[phone]E-mail stephane.cote_at_space.gc.ca
2
RADARSAT-1 Image of Montreal
CSA HQ
Mount Royal
Downtown
Images processed by the Canadian Data Processing
Facility (CDPF)
3
Mission Status General
Extended Mission
Nominal Mission Lifetime (5.25 years)
2004
2003
2002
2001
2000
1999
1998
1997

• 1996

1995
Launch (11-95)
Commissioning (04-96)
Routine Operations
Calibration Phase (09-97) TNRL established
Maintenance Phase
HS2 Failure (AMM, year 1) Switched to HS1
Primary momentum wheels (10-99) Speed variations
high temp. Switched primary to redundant
OBR1 high error (04-02) Switched to OBR2
Redundant pitch wheel (12-02) Alternate
procedure S/C operations uses torque rod inputs
4
Mission Status SAR Imagery
Mission designed lifetime 5.25 years Since
launch 8.5 years

• Radiometric recalibrations since 1998
• Once W2, S1, S2, S4, S6
• Twice W1, F4, S3, S7, EL1
• Unaffected beams since Cal. Phase
• W3, S5, F1, F2, F3, F5

• As end of Dec. 2003
• Total orbit completed 42 549
• Total number of customer image requests 166 000
• Total products (image frames) generated 354 000
• Total minutes of data 314 000
• Total OBR minutes 67 576
• Average System Performance (FY2003-2004) 95.7

N. of successful receptions

• N. of requests planned

5
RADARSAT-1 IQ and Calibration Plan

• Point Target Monitoring
• Spatial Resolution
• Focusing and Energy Concentration in Main Lobe
• Geometrical Integrity

• Distributed Target Monitoring
• Image level of any point within covered area,
  relative to normalized scattering cross-section

RSAT-1 Precision Transponders (RPTs) Average of
16 acquisitions/cycle 1357 to date

• Amazon Rainforest
• Average of 13 acquisitions/cycle
• 987 to date

• Impulse Response Width (IRW)
• Peak Side-Lobe Ratio (PSLR)
• Integrated Side-Lobe Ratio (ISLR)
• Absolute Location Error (ALE)

Radiometric Accuracy Profiles so backscatter
coefficients ?o bo brightness parameters
Ground Processor Calibration Workstation
6
Point Target MonitoringRADARSAT-1 Precision
Transponder Sites
7
Image Quality History of CDPF Products

• Results since Early Maintenance Phase (97-98)
• Stable through Extended Mission Phase (early
  2001)
• Point Target Parameters
• Impulse Response Width (IRW) ? Spatial
  Resolution
• Peak Side Lobe Ratio (PSLR) ? Focusing
• Integrated Side Lobe Ratio (ISLR) ? Energy
  Concentration in main lobe
• Absolute Location Error (ALE) ? Geometrical
  Registration

8
Image Quality Historyof CDPF Products
Spatial Resolution (IRW)
Range Azimuth IRW Fine Beams (BW30.00 MHz)
Range Azimuth IRW S1, S2 and EL1 Beams
(BW17.28 MHz)
Range Azimuth IRW S3-S7, Wide and EH Beams
(BW11.58 MHz)
9
Image Quality Historyof CDPF Products
Focusing (PSLR) All Beams
Energy Concentration in main lobe (ISLR) All Beams
Increased solar activity
Geometrical Registration (ALE) All Beams
10
Distributed Target MonitoringAmazon Rainforest
11
Radiometric Calibration

• Purpose
• Standardize the image intensity response across
  instances of a product
• Measure changes in target reflectivity over time
• Compare different targets in one image or between
  two images
• Radar beams are calibrated to bo (brightness
  coefficient)
• 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

12
Radiometric Calibration
Radiometric Accuracy Beams F5 Never recalibrated
Radiometric Accuracy Beam W1 Recalibrated twice
Elevation Beam Pattern for W1
13
ScanSAR Radiometric Calibration and Image Quality

• ScanSAR modes are aggregate of multiple beams
  acquired together to generate a single, larger
  image
• From 300X300 km to 500X500 km
• CDPF ScanSAR products
• SCNA composite of W1, W2
• SCNB composite of W2, S5, S6
• SCWA composite of W1, W2, W3, S7
• SCWB composite of W1, W2, S6
• Radiometrically calibrated to ?o effective Feb.
  1, 1999
• ScanSAR processor in CDPF major upgrade in 2002
• Remarkable improvements in image quality
• Elimination of processing problems
• Scalloping
• Location Error due to PRF Ambiguity
• Visibility of Beam Boundary
• Automatic Gain Control (AGC) Saturation Error

14
ScanSAR Radiometric Calibration
Gamma nought (brightness) for SCNB Ascending
Pass, 18-Jan-2004 (M0356804)
? o ?o tan(incidence angle)
15
ScanSAR Image Quality Scalloping
16
ScanSAR Image Quality Beam boundaries
17
ScanSAR Image Quality 2002 Processor Upgrade
Baltimore
Washington
RADARSAT-1 ScanSAR Wide B Orbit
33098 08-MAR-2002 Eastern United States
Cape Hatteras
Cape Hatteras
18
Boreal Forest Radiometric Experiment

• Amazon images used for calibration are recorded
• Aging considerations for the On-Board Recorder
  (OBR)
• Alternative sites within Canadian mask are being
  explored to support radiometric analyses
• Site Survey
• Canadian Tundra/Taïga belt, Boreal Forest
  locations
• Preliminary results
• Most homogeneous site, near Hearst, Ontario,
  chosen for testing
• Radiometric test method
• Developed by adapting existing methodology for
  Amazon
• Average of 10 acquisitions per cycle - total of
  200 to date
• Entire range of incidence angles covered
• Use of W1, W2, W3, S7 and EL1 beams

19
Boreal Forest Hearst Site
20
Boreal Forest Hearst Site
21
Boreal Forest Measurement Campaign

• Started in January 2003
• Assessment of seasonal variations
• Local weather data adjoined to radiometric
  analyses
• Investigate possible correlations between gamma
  nought (radar brightness) and
• Snow on the ground - winter
• Humidity level - summer
• Current precipitation conditions
• Experiment on-going
• Encouraging results

22
Boreal Forest Measurement Campaign
Pattern comparison for M0363850 Beam W1,
Ascending, 23-March-2004
(Amazon)
23
Image Quality in Degraded Attitude Determination

• RADARSAT-1 spacecraft currently uses Horizon
  Scanner 1 (HS1)
• Spacecraft is in 'Attitude Determination Mode 1'
  (ADM1)
• In October 2000, concerns began to rise of the
  possibility of HS1 failure
• Would result in operating the spacecraft in a
  mode known as 'Attitude Determination Method 3'
  (ADM3)
• ADM1 Sun Sensor Horizon Scanner
• ADM3 Sun Sensor Magnetometer
• ADM3 reduced attitude control performance of
  spacecraft compared to ADM1

24
ADM3 Experiments

• Operating the spacecraft in ADM3 mode
• Between January and December 2001 experiments
  were conducted to better understand the impact on
  processing and image quality
• 69 images were acquired at multiple orbits over
  three sites
• The ADM3 flight test series have been a
  significant success
• Images processed successfully with good image
  quality and calibration performance
• More re-work on images is expected at CDPF in
  ADM3 mode
• Due to increased error in estimation of Doppler
  ambiguity

25
ADM3 Experiments
F4 Ascending over Fredericton
ADM3
ADM1
9 Dec 2001
15 Nov 2001
26
Conclusions

• Image Quality point target measurements indicate
  that IQ is still maintained better than system
  specification
• Radiometric Calibration routinely monitored and
  maintained
• Where required, re-calibration is performed to
  accommodate changes in SAR antenna elevation beam
  pattern
• Internal Calibration has remained stable since
  1999
• Significant improvements have been made in image
  quality and radiometry of ScanSAR with upgrades
  of CDPF in 2002
• After being launched in 1995 and declared
  operational on April 1, 1996, RADARSAT-1
  performance continues to be excellent in extended
  mission
• System aging considerations have led to
  successful experiments
• Hearst radiometric measurements (OBR)
• ADM3 experiments (HS1)