Airborne Campaigns for Pol-InSAR Applications Development

Transcription

1 Airborne Campaigns for Pol-InSAR Applications Development Irena Hajnsek, Alberto Moreira, Malcolm Davidson German Aerospace Center European Space Agency Folie 1 irena.hajnsek@dlr.de Outline Why do we need airborne campaigns? First airborne demonstration of Pol-InSAR What is the status of application development with Pol-InSAR? Application Agricultural Application Ice and Snow Application Urban Area Application Contribution of airborne campaigns to spaceborne Folie 2 1

2 Why do we need airborne campaigns? Innovation Specifications needed for future satellite sensors Test advanced imaging modes (Pol-InSAR, digital beamforming, etc.) Development Development of algorithm for quantitative parameter estimation Development of new application products Observations to with which to calibrate or validate satellite retrievals Data Availability Detailed information's in critical areas Key information that cannot currently be measured from space Young researcher education and preparation to satellite SAR sensors DLR s aircrafts Requirements to an airborne system Flexible and modular SAR system System availability Complete operational processing chain Fast data delivery High data quality Folie 3 First airborne Pol-InSAR campaigns Airborne System/ Institution First Pol- InSAR demo Frequency Test site Application JPL 1993 C-,L-& Portage Lake Maine DNSC (DTU) 1995 C-band Foulum DLR-HR 1998 OP ONERA 1999/2004 X-/ L-& Bretigny/Le Landes Agri/ JPL 2000 Camp Lejune JAXA 2000/2001 Tutori/Tsukuba Intermap 2000/2002 Capitol DSTO 2005 (?) X-band Australia Folie 4 2

3 AIRSAR s first Pol-InSAR JPL, USA C-band Interferograms Date: 1993 Correlation Objectives: To measure attenuation at different frequencies in the canopy Understand temporal decorrelation in the canopy See how the interferometric phase varied with polarimetric decompositions Summary: Airborne InSAR experiments have shown good correlation at GeoSAR: Dual Campaign in Camp Lejune NC USA 2000 Folie 5 First airborne Pol-InSAR campaigns Airborne System/ Institution JPL First Pol- InSAR demo 1993 Frequency C-,L-& Test site Portage Lake Maine Application DNSC (DTU) 1995 C-band Foulum DLR-HR 1998 OP ONERA 1999/2004 X-/ L-& Bretigny/Le Landes Agri/ JPL 2000 Camp Lejune JAXA 2000/2001 Tutori/Tsukuba Intermap 2000/2002 Capitol DSTO 2005 (?) X-band Australia Folie 6 3

4 EMISAR s first Pol-InSAR DNSC, Denmark Test site: Foulum, Denmark Campaign purpose: Crop classification Acquisition date: May 3, 1995 Acquisition/processing: Technical University of Denmark Frequency: C-band EMISAR DEM Coherence magnitude (RGB-coded Pauli basis) Folie 7 First airborne Pol-InSAR campaigns Airborne System/ Institution JPL First Pol- InSAR demo 1993 Frequency C-,L-& Test site Portage Lake Maine Application DNSC (DTU) 1995 C-band Foulum DLR-HR 1998 OP ONERA 1999/2004 X-/ L-& Bretigny/Le Landes Agri/ JPL 2000 Camp Lejune JAXA 2000/2001 Tutori/Tsukuba Intermap 2000/2002 Capitol DSTO 2005 (?) X-band Australia Folie 8 4

5 E-SAR s first Pol-InSAR DLR, Germany Interferometric Coherence in Three Polarisations Interferometric Phase in Three Polarisations HH Inversion/ Validation Oberpfaffenhofen D- Representation of Tree Height Folie 9 First airborne Pol-InSAR campaigns Airborne System/ Institution First Pol- InSAR demo Frequency Test site Application JPL 1993 C-,L-& Portage Lake Maine DNSC (DTU) 1995 C-band Foulum DLR-HR 1998 OP ONERA 1999/2004 X-/ L-& Bretigny/Le Landes Agri/ JPL 2000 Camp Lejune JAXA 2000/2001 Tutori/Tsukuba Intermap 2000/2002 Capitol DSTO 2005 (?) X-band Australia Folie 11 5

6 RAMSES s first Pol-InSAR ONERA, France σ Αrg(γ) Date: Sep 1999 Campaign: ORANGE Objective: Polarimetry studies Test site: Bretigny, France Frequency: X band Imaging mode: Single pass γ Folie 12 First airborne Pol-InSAR campaigns Airborne System/ Institution First Pol- InSAR demo Frequency Test site Application JPL 1993 C-,L-& Portage Lake Maine DNSC (DTU) 1995 C-band Foulum DLR-HR 1998 OP ONERA 1999/2004 X-/ L-& Bretigny/Le Landes Agri/ JPL 2000 Camp Lejune JAXA 2000/2001 Tutori/Tsukuba Intermap 2000/2002 Capitol DSTO 2005 (?) X-band Australia Folie 14 6

7 PI-SAR s first Pol-InSAR JAXA, Japan First experiments Objective Oct. 8, 2000 March 3, 2001 Tree height, InSAR capability (~90m Bp) Nov Surface deformation with control flight (5m tube) Test site Frequency Tottori, Tsukuba 1.27GHz Mt. Fuji 1.27GHz Folie 15 First airborne Pol-InSAR campaigns Airborne System/ Institution First Pol- InSAR demo Frequency Test site Application JPL 1993 C-,L-& Portage Lake Maine DNSC (DTU) 1995 C-band Foulum DLR-HR 1998 OP ONERA 1999/2004 X-/ L-& Bretigny/Le Landes Agri/ JPL 2000 Camp Lejune JAXA 2000/2001 Tutori/Tsukuba Intermap 2000/2002 Capitol DSTO 2005 (?) X-band Australia Folie 16 7

8 TopoSAR first Pol-InSAR Intermap, Germany TopoSAR (formerly AeS-1) Simultaneous X-band HH and quad-pol bandwidth: MHz centre frequency First Pol-InSAR demonstration in 2000 First operational Pol-InSAR campagain in July 2002: Capitol (Washington State, USA) for the purpose of bald-earth DTM generation beneath forest canopy Gulfstream Commander with X- and antennas Capitol example: Magnitude (HH) Aerial Photograph Pol-InSAR derived DTM Folie 17 Vegetation Height Biomass Canopy Extinction Structure Underlying Topography Ecology Management Ecosystem Change Carbon Cycle Agricultural Vegetation Soil Moisture Content Soil Roughness Height of Vegetation Layer Extinction of Vegetation Layer Moisture of Vegetation Layer Farming Management Water Cycle Desertification Ice and Snow Topography Penetration Depth Depth of Snow-Ice Layer Extinction Density of Ice Bodies Ecosystem Change Water Cycle Water Management Urban Areas Geometric Properties Dielectric Properties Urban Monitoring Folie 18 8

9 First Pol-InSAR Demonstration: Application SIR-C/X-SAR / Test Site: Kudara, Russia / Pauli RGB 1994: SIR-C / X-SAR acquires the first POL-InSAR data set 1996: First publication on Pol-InSAR. 1998: First Pol-InSAR forest height estimation. Folie 19 E-SAR/Test Site: Oberpfafenhoffen Azimuth Range Baseline=25m HH HV VV Baseline=40m HH HV VV Folie 20 9

10 Airborne Campaigns with applications Campaign Objectives: E-SAR campaigns over different forest types in Europe for forest height retrieval: 1. Algorithm improvement 2. Model validity 3. Validation 4. Product development Three main forest types have been investigated: Temperate forest Boreal forest Mediterranean forest Fichtelgebirge / Germany Corridor / Spain Bayrischer Wald / Germany Traunstein / Germany Oberpfaffenhofen / Germany Folie 21 Airborne applications Biomass [t/ha] Pic.abies Pi.silv. Ab.picea Pseu.menz Lar. dec Gg.sylv. Querc.rab. Frax.exc. Bet.pend. Rob.pseu. Por.rob Height (in Situ) [m] Height [m] Folie 22 10

11 Tropical forest applications Campaign Objectives: 1. Investigation of L- and over tropical forest for forest height estimation (visebility of the ground, validity of the RVoGM etc.) 2. Is there a relation of forest height and forest biomass over tropical forest? 3. Is there an empirical relation between the radar backscatter and the forest biomass using L- and/or? Kalimantan - Indonesia Folie 23 MAWAS: Tropical forest height RGB Image X/L/P Band rg az Tropical Height from Pol-InSAR Pol-InSAR S HH SHV [ S] 1 = SVH SVV S HH SHV [ S 2 ] = SVH SVV z bl 1 2 n r height over tropical forest in Indonesia x Folie 24 h y 11

12 Preliminary results: Tropical forest applications 1. Both frequencies L- & are suitable for forest height estimation 2. Research is still ongoing and a definite answer is pending relation between height/biomass 3. Six different definitions of biomass have been found in the literature and have been used for comparison. No significant relation between biomass and radar backscatter Top height of transects : 26m were found ( t/ha) Plot design of ground truth: 20 squares of 10m x 10m in two transects L- against derived heights 45 whole scene ground plot H100[m] P Band (blue) L Band (red) Boreal Helsinki Area See presentation by Florian Kugler Folie 25 Cooperation with HUT Campaign Objectives: 1. Investigation of Pol-InSAR in over boreal forest areas. 2. Comparison of Pol-InSAR derived forest heights with HUT-SCAT data HUT-SCAT Profile Folie 26 12

13 different frequencies Short summary: 50% HH & VV 1. Significant correlation between & HUT- SCAT derived forest heights X-band 2. Significant correlation between X-band & HUT- SCAT derived forest heights ~ r ~ r γ m( w) γ ( V + γ ~ r ( w ) = exp( iφ w) = exp( iφ0 ) r 0 ) γ~ V 1+ m( 30º C-band & HUTSCAT Height Profiles X-band & HUTSCAT Height Profiles Hutscat look E-SAR / Test Site: Helsinki, Finland Folie 27 Vegetation Height Biomass Canopy Extinction Structure Underlying Topography Ecology Management Ecosystem Change Carbon Cycle Agricultural Vegetation Soil Moisture Content Soil Roughness Height of Vegetation Layer Extinction of Vegetation Layer Moisture of Vegetation Layer Farming Management Water Cycle Desertification Ice and Snow Topography Penetration Depth Depth of Snow-Ice Layer Extinction Density of Ice Bodies Ecosystem Change Water Cycle Water Management Urban Areas Geometric Properties Dielectric Properties Urban Monitoring Folie 28 13

14 E-SAR s Agricultural Campaign in Tunesia 2005 Airborne Campaign: AQUIFEREX within AQUIFER support the national authorities and international institutions with Earth Observation (EO) based technology to better manage internationally shared water resources and aquifers strengthen overall and integrated water management practices build-up of an independent service provision capacity to ensure local service delivery after the project cycle to achieve the longer-term goal of service sustainability Campaign Execution November 2005 South Tunesia (Ben Gardane,Gabes Region) Important observation Parameter: Land-cover and use & soil moisture E-SAR: C-band quad-pol E-SAR: quad-pol AVIS (rgb=659, 550, 477 nm) Folie 29 Bare Surfaces: Soil Moisture AQUIFEREX Pol. Coh. Matrix X-Bragg Bragg Roughness Term T11 T12 T21 T22 T33 T11 T12 = T21 T22 + T11 T22 T33 E-SAR / Test Site: Aquiferex, Tunisia sensor scattering plane surface β LOS / Pauli RGB Dielectric Constant Vol. Soil Moisture Folie 30 14

15 Alling/Germany 2000 Test Site: Kuettighoffen, Switzerland V SAR Interferometric Coherence: ~ r γ ( w) = I I 0 ~ r γ( w ) = exp( iφ hv r 2 σ( w) z' I = exp( iκ zz')exp dz' 0 cosθ 0 hv r 2 σ( w) z' I0 = exp dz' cosθ D Height Map ~ r r γv ( w ) + m( w ) ) 1 + m( w ) 0 r E-SAR / Test Site: Alling, Germany 40 db 0 db Differential Extinction Folie 31 AGRISAR: Soil Surface and Crop Parameter (April-Aug.2006) 1. Building up a data base for agricultural parameter estimation over a whole vegetation growth period - has been started mid April support the space segment activities at ESA with respect to the Sentinel Program answering open questions concerning system constellation (single, dual, qual polarisation, revisit time, etc) 16 Radar data acquisition flights during three months Soil Moisture Station Polarimetry Frequency X-band interferometry Bowen-Ratio Station Laser-Profiler Folie 32 15

16 for 3 intensive periods 19/04/06 06/06/06 05/07/06 RGB Pauli Folie 33 Vegetation Height Biomass Canopy Extinction Structure Underlying Topography Ecology Management Ecosystem Change Carbon Cycle Agricultural Vegetation Soil Moisture Content Soil Roughness Height of Vegetation Layer Extinction of Vegetation Layer Moisture of Vegetation Layer Farming Management Water Cycle Desertification Ice and Snow Topography Penetration Depth Depth of Snow-Ice Layer Extinction Density of Ice Bodies Ecosystem Change Water Cycle Water Management Urban Areas Geometric Properties Dielectric Properties Urban Monitoring Folie 34 16

17 Snow First Pol-InSAR Snow Experiment in Austria 2004 E-SAR: Kuehtai / Austria 2004 Cooperation with University of Innsbruck Campaign Objectives: Investigation of Pol-InSAR for snow characterisation Snow appears as a Volume / HH Image X-band DEM HH-HH Coherence / Baseline 1 HH-HH Coherence / Baseline 2 Folie 35 Snow θ o z 0 + h v θ r n = sin( θ 0 ) sin( θ r ) z 0 Snow depth can be Snow potentially estimated Baseline 1 (20M): Δφ=17 -> Δz=1.48m Baseline 2 (40M): Δφ=28 -> Δz=1.22m / HH Image HH-HH Coherence XX-XX Coherence VV-VV Coherence Folie 36 17

18 SVALEX 2005 Svalbard Airborne Experiment DLR-HR () * Pol-InSAR measurements over land ice * penetration depth at different frequencies * retrieval of glacier topography * mapping of internal ice structures AWI-Potsdam (Alfred-Wegner Institute) * measurements of optical, chemical, and * physical properties of Arctic aerosols, in particular Arctic haze AWI-Bremerhaven (Alfred-Wegner Institute) * boundary layer meteorology * transfer of atmospheric momentum to sea ice * use of radar imagery to quantify ice ks Sensors: 1. E-SAR Radar System 2. Meteopod (wind velocity, humidity, temp.) 3. Laser Altimeter 4. KT15 (surface temp.) 5. Linescan camera AWI-DLR Team Sea/Land ice SAR data acquisition Folie 37 SVALEX 2005 Svalbard Experiment 05 Sea Ice ( quad pol) Airborne Radar Campaign at Spitzbergen over Sea Ice in X- & quad pol over Land Ice (Glacier) in Pol-InSAR L- and azimuth Land Ice range Land Ice Folie 38 18

19 SVALEX 2005 Svalbard Experiment 05 Preliminary Summary Investigation of penetration depth at different frequencies, where the corner reflector are used as a reference point. First results show that lower penetration occurs as expected in X-band, and. Implementation and modification of coherent scattering models for the characterisation of the ice volume and Pol- InSAR analysis at different frequencies showed good agreement with the observables. Corner Summit Sea Ice & Meteopod Meteo Station See presentation by Jayanti Sharma Folie 39 Vegetation Height Biomass Canopy Extinction Structure Underlying Topography Ecology Management Ecosystem Change Carbon Cycle Agricultural Vegetation Soil Moisture Content Soil Roughness Height of Vegetation Layer Extinction of Vegetation Layer Moisture of Vegetation Layer Farming Management Water Cycle Desertification Ice and Snow Topography Penetration Depth Depth of Snow-Ice Layer Extinction Density of Ice Bodies Ecosystem Change Water Cycle Water Management Urban Areas Geometric Properties Dielectric Properties Urban Monitoring Folie 41 19

20 Campaigns over Urban Area over Munich 2005/6 Campaign Objective: 1. Identification of Coherent Scatterers at different Frequencies 2. Characterisation of Coherent Scatterer (dielectric and geometric properties) Dihedral Size: Vertical: 80 x 80 cm Horizontal: 80 x 80 cm Orientation: 1th dihedral: 0 2nd dihedral: 5 Ascending Munich Acquisition: 13. Oktober 2005 Descending Tracks: 0, 30, 90, 270m 3 independent Folie 42 Urban Areas: Coherent Munich 2005/6 E-SAR / Test Site: Munich, Germany : First CS s dedicated experiment R: HH-VV G: HV B: HH+VV R: G: C-band B: X-band Folie 43 20

21 Coherent Scatterers Red: Dihedrals Green: Dipoles Blue: Flat Plates Amplitude Entropy Alpha angle See Microwaves presentation and Radar by Institute Kostas Papathanassiou & Luca Marotti Folie 44 Looking Forward Development of parameter estimation methodology / algorithms: Snow / Ice / Information product definition and validation / Projection of product spec s onto system design. Exploration of new and innovative observation spaces: From E to F ( SAR ) From S to Q ( Pol ) From R to S ( Pass ) From M to B ( Static ) ALOS / Pal-SAR TerraSAR-X RadarSAT-2 Tandem-X Folie 45 21