MATISSE : version 1.4 and future developments

Transcription

1 MATISSE : version 1.4 and future developments Advanced Earth Modeling for Imaging and the Simulation of the Scenes and their Environment Page 1 Pierre Simoneau, Karine Caillault, Sandrine Fauqueux, Thierry Huet, Jean Claude Krapez, Luc Labarre, Claire Malherbe, Christophe Miesch Office National d Études et de Recherches Aérospatiales

2 MATISSE-v1.4 : 4 computation modes «Imaging» mode LOS mode Direct computation API Radiance and Transmission along a LOS / σ = 5 cm -1 Radiance and transmission Images / σ = 5 cm -1 HSR mode (LOS) High spectral resolution thermal radiance and transmission Page 2

3 Contributors Atmospheric variability Atmospheric extinction Atmospheric emission and scattering Cloud emission and scattering Observed radiance Ground emission and reflection

4 Atmospheric emission and scattering Atmospheric emission and absorption CK model cm -1 (3 13 µm) / δσ = 5 cm -1 / step 5 cm -1 Atmospheric multiple scattering DOM Atmospheric data thermodynamic data 1D profiles database : AFRL + TIGR (1760 profiles) 2D profiles database : 2D climatology 3D profiles scene atmospheric variability User atmospheric profile aerosols data GADS climatology database (D Almeida & al) aerosols variability AFRL aerosols : Rural / Urban / Fog / Maritime / Tropospheric AP (DRDC collaboration) : Marine Boundary Layer Aerosols

5 Clouds generation and radiance Two options 1/ Partial coverage of Sc clouds (imaging mode) cloud shape generator IPA + (BRDF, BTDF, ε) cloud radiance spatial variability (but time consuming) Option : radiance texture model (PSD) 2/ Total coverage of Sc or Ci clouds (no horizontal variability) direct radiative computation

6 Ground data sets and radiance DTED 30 global coverage 3 Europe Land use / (r,ε) global 30 (IGBP-DISCover + ASTER) 1D Ground thermal model Periodic radiative energy deposit Fourier Method Solar energy deposit : 2 streams model 305 Température K ASST (sea temperature) 293

7 LOS modes Standard mode Spectral radiance and transmission (or integrated with apparatus function) cm -1 (3 13 µm) / δσ = 5 cm -1 / step 5 cm -1 / CK model Uniform clouds : Sc or Ci Uniform clouds : Sc or Ci Refraction No refraction Elipsoïde DTED and (no TO DTED) properties User T ground and albedo θ Atmospheric Atmospheric variability variability Page 7

8 LOS modes Propagation in the Marine Boundary Layer Maritime aerosol model AP (DRDC) Surface Data. P air, T air, T mer, HR, V moy, fetch, visi, Ma, H vagues, Radio-sounding. Temp / American MBL 30 m AP Thermodynamic profile Aerosols profile Cn2 profile : Sea Surf Zmax ~ 3 km Extrapolation : Thermo. Profile + climatology AP Aerosols profile + AFRL Sea Surf TOA High accuracy path modeling Multiple path propagation (DRDC)

9 High spectral atmospheric thermal radiance and transmission along a LOS cm -1 (3 13 µm) / δσ = cm -1 2 independent modes : all molecules all molecules excepted H 2 O and CO 2 Refraction All MATISSE thermodynamic profiles No aerosols (including 3D profiles) No ground radiation Page 9

10 Computers Workstations SUN / Solaris 2.8 IBM / AIX 4.3 PC Windows (2000 / XP) Full version Light version (1 DVD) limited functionalities LOS only (CK and HSR) No large databases (no GADS / no 3D thermodynamic scene / no DTED)

11 Results

12 Spectral radiance image W/m 2.sr.cm Geostationary satellite 15/06/ h00 UTC US Std / Rural 23 km long. 2. W Nadir angle : 7 FOV : 1.2 x x 200 pixels Multiple scattering on.029 Total computation time for 1 wavelength (1250 cm -1 ) ~ 34 Sun Workstation Thermal model computation time ~ 30 Sun Workstation (# ground facets ~ )

13 Integrated over sensor bandwidth radiance images W/m 2.sr 1.86 W/m 2.sr Waveband 3 5 µm ~ 6.5 h Sun Workstation (260 spectral elements) Waveband 8-12 µm ~ 2.5 h Sun Workstation (84 spectral elements) 24

14 Future works: MATISSE-v2.0 (end 2008)

15 MATISSE-v2.0 : main functionalities High spatial resolution sea irradiance images finest spatial resolution (footprint) = 1m sub-metric variability of the radiance solar glint effect irradiance computation for all background types multi-resolution approach Page 15

16 Sea radiance modeling Multi-resolution problem L L L L Low spatial resolution (L >> Lc) Moderate spatial resolution (L >> Lc) High spatial resolution ( 1m) Statistic approach ε(λ,θ,ϕ), ρ(λ,θ,ϕ,θ,ϕ ) Require : <m> and σ Semi-statistic approach : ε(λ,θ,ϕ), ρ(λ,θ,ϕ,θ,ϕ )

17 1/ Low spatial resolution. <m> = 0. Variance: σ = ε(λ,θ,ϕ) ρ(λ,θ,ϕ,θ,ϕ,ϕ ) Pixel en y Pixel en y Pixel en x Pixel en x

18 2/ Moderate and high spatial résolution Computation of ε(λ,θ,ϕ) and ρ(λ,θ,ϕ,θ,ϕ )semi-statistic approach Capillarity waves : <m> = 0 / σ 2 capillarity Gravity waves : <m> over h and γσ 2 given by γ 2 ways 1/ Spatial resolution fixed = 1m FFT surface generation + repetition computation of ρ(<m>,σ) and ε(<m>,σ) with <m> and σ computed in each facet (1mx1m) 2/ Spatial resolution = footprintsuperposition of 2D sinusoids computation of ρ(<m>,σ) and ε(<m>,σ) with <m> and σ computed in each facet with dimension ~ footprint

19 Radiance image (patch method) PROTO 315 m Radiance Relief image image (BRDF+ thermal emission) 315 m 256x256 pixels Wind speed = 9 m.s-1 No shadow 500 m 45 45

20 Comparison 1D analytical BRDF with reference model θ sun = m 4 m u 10 =5m/s u 10 =10m/s u 10 =15m/s L=64 m L=4 m

21 Comparison 1D analytical BRDF with reference model θ sun = -80 u 10 =5m/s u 10 =10m/s u 10 =15m/s L=64 m L=4 m

22 Comparison 1D analytical emissivity with reference model u 10 =5m/s u 10 =10m/s u 10 =15m/s L=64 m L=4 m

23 Solar glint effect PROTO Ω Θ Θ 512*512

24 Summary MATISSE-v1.4 Imaging mode / LOS mode / HSR mode / API mode SUN /IBM / PC Windows MATISSE-v2.0 (2008) Sea images (high spatial resolution) Irradiance computation MATISSE-v1.5? (2007) Fast multiple scattering computation Improvement thermal model computation visible