Program Monthly Review RADARSAT-2 TOPS Implementation of TOPS mode on RADARSAT-2 in Support of the Sentinel-1 mission G. Davidson, V. Mantle, B. Rabus, D. Williams MDA D. Geudtner ESA 1 MDA Proprietary
RADARSAT-2 TOPS Background • Sentinel-1 Interferometric Wide Swath (IW) is main mode of operations, using TOPS technique • Experimental RADARSAT-2 TOPS mode resembles as closely as possible the performance of the Sentinel-1 IW mode • Processing of RADARSAT-2 TOPS data in the Sentinel-1 L1 (SLC) data product format • Provide Sentinel-1 like RADARSAT-2 TOPS SLC data to Copernicus services and users prior to the Sentinel-1 launch • RADARSAT-2 operates at same C-band frequency • Phased array antenna supports multiple modes • Provide data as similar as possible to Sentinel-1 • MDA develops the Sentinel-1 Instrument Processing Facility (IPF) under ESA contract MDA Proprietary
RADARSAT-2 TOPS TOPS Review azimuth burst Scan rate k q sub-swath 1 focused burst sub-swath 2 sub-swath 3 range Targets swept by entire beam • illuminated by entire azimuth antenna pattern – no scalloping • illumination time reduced, and resolution reduced by a factor 𝛽 = 1 + 𝑙 𝜄 𝑆 𝑤 • Doppler sweep rate 𝑙 𝑡 = 2𝑤 𝑡 𝜇 𝑙 𝜄 MDA Proprietary
RADARSAT-2 TOPS RADARSAT-2 TOPS Beam Design • Based on existing RADARSAT- 2 ‘ ScanSAR Narrow B’ mode • Mode comparison: MDA Proprietary
RADARSAT-2 TOPS TOPS Implementation • TOPS: forward electronic scanning of antenna in azimuth • RADARSAT-2: antenna steering in coarse steps – Beam switching time • In-orbit experiment found minimum time to be about 15 ms – Number of Timing Control patterns • Limits total number of azimuth steps across all sub-swaths • Max azimuth pointing angle (burst overlap) vs. step size MDA Proprietary
RADARSAT-2 TOPS TOPS Implementation MDA Proprietary
RADARSAT-2 TOPS Expected effect of coarse step size • Modulation of raw signal amplitude from point target – Increased sidelobes in point target response (approx 0.7 dB degradation) – Slight variation in radiometry (approx 0.05 dB) Simulation results MDA Proprietary
RADARSAT-2 TOPS Analysis of Expected RTOPS Coherence • Analyze spectral shifts due to burst mis-synchronization and difference in antenna pointing • Statistical analysis of 𝛽 − 1 RADARSAT-2 InSAR stacks Δ𝑔 𝑈 = 𝑙 𝑏 𝑈 𝑒𝑓𝑚 𝛽 • Probability distribution of coherence due to burst synchronization, Doppler 𝑔 = Δ𝑔 centroid difference, and 𝑒𝑑 Δ𝑔 𝛽 baseline MDA Proprietary
RADARSAT-2 TOPS Analysis of RADARSAT-2 InSAR stacks • Statistical characterization of baseline, burst mis- synchronozation and Doppler centroid difference 50 Percentile (pairwise 90 Percentile (pairwise absolute difference) absolute difference) Burst Synchronization 16 milli-seconds 41 milli-seconds Doppler Centroid 36 Hz 92 Hz Variation Perpendicular Baseline 160 meters 435 meters MDA Proprietary
RADARSAT-2 TOPS Analysis of RADARSAT-2 InSAR stacks PDFs for the pairwise differences: • Baseline • Doppler centroid • Start time synchronisation Create 3-D PDF for coherence: 𝛿(𝑐 ⊥ , 𝑈 𝑒𝑓𝑚 , Δ𝑔 𝑒𝑑 ) Integrate 3-D PDF to get probability of coherence exceeding a certain value 50% chance of exceeding 0.84 90% chance of exceeding 0.66 MDA Proprietary
RADARSAT-2 TOPS Processing • Pre-processing with RADARSAT-2 processor – Ingest – Range compression – Formatting • Doppler centroid estimation • Sentinel-1 Instrument Processing Facility (IPF) – Inputs range compressed data and formatted annotation – upsampling – RCMC and SRC – Azimuth compression – Resampling – Formatting in Sentinel-1 L-1b (SLC) product format MDA Proprietary
RADARSAT-2 TOPS Acquisition of various Scene Types • Sea-ice (Lancaster Sound, Canadian Arctic) • CSA Transponder sites, Montreal and Ottawa (calibration) • Ocean Currents (Gulf stream, Agulhas current) • Gibraltar (ship detection) • Richmond, BC, Canada • Amazon, Brasil (calibration) • Pacific Doldrums (NESZ estimation • Markermeer, Netherlands (NESZ estimation) • Uyuni Salt Flats, Bolivia, (InSAR pair) • Petermann Glacier (Greenland) (InSAR pair) • Lambert Glacier (Antarctica) (InSAR pair) • Mt. Etna Volcano (Italy) (InSAR pair) • Mexico-City (Mexico) (InSAR stack) 12 MDA Proprietary
RADARSAT-2 TOPS Richmond Acquisition overlap overlap Slant range images MDA Proprietary
RADARSAT-2 TOPS Richmond Acquisition GRD product MDA Proprietary
RADARSAT-2 TOPS Image Quality Results • Point target analysis – target of opportunity in Richmond scene – Azimuth response shape agrees with simulation MDA Proprietary
RADARSAT-2 TOPS Image Quality Results • Point target analysis – CSA Transponders in Montreal scene – Expected values in range are from Hamming (0.75) window – Expected values in azimuth are from simulation of effect of coarse step St. Hubert Ottawa Expected Range Resolution [m] 5.0 5.06 5.09 Range ISLR [dB] -16.1 -15.62 -15.37 Range PSLR [dB] -21.3 -21.19 -20.44 Azimuth Resolution [m] 29.4 / 29.2 / 29.0 29.08 29.50 Azimuth ISLR [dB] -15.39 -15.37 -15.51 Azimuth PSLR [dB] -20.66 -21.62 -20.67 MDA Proprietary
RADARSAT-2 TOPS Image Quality Results • Radiometry over Amazon – Apply gain factors so that g 0 over Amazon is approx. -6.5 dB – g 0 profile for three sub-swaths vs. incidence angle MDA Proprietary
RADARSAT-2 TOPS Image Quality Results • Azimuth profile over a burst – Verify removal of residual scalloping using model of antenna element pattern sub-swath 1 sub-swath 2 sub-swath 3 MDA Proprietary
RADARSAT-2 TOPS Image Quality Results • Noise Equivalent Sigma 0 (NESZ) – Estimate using HV data over Doldrums – Three sub-swaths vs. incidence angle, compared to expectation MDA Proprietary
RADARSAT-2 TOPS Doppler Grid • Gulf Stream Scene 96 Doppler estimates vs. range over all 3 sub-swaths – 32 overlapping range segments per burst • 5 estimates in azimuth per burst – Deramp in azimuth to remove antenna scanning Superimposed on ocean effect current map (CLS using – Overlapping azimuth ASAR 2008) segments for Doppler estimates – Correct for small error in Agulhas Scene Doppler sweep rate • Deviation from mean Doppler centroid estimate derived from attitude information MDA Proprietary
RADARSAT-2 TOPS Interferometry Uyuni : Sub-swath 1, Burst 1 coherence phase phase superimposed on amplitude MDA Proprietary
RADARSAT-2 TOPS Conclusion • Experimental TOPS mode successfully implemented on RADARSAT-2 – Effect of coarse step size is not severe • Statistical analysis of the timing and antenna pointing of existing InSAR stacks shows good probability of coherence • Image quality results in line with performance expectations • Demonstration of Doppler grid over ocean currents • Good coherence observed in interferograms MDA Proprietary
Recommend
More recommend