Compact Polarimetry Potentials

My-Linh Truong Loï, Jet Propulsion Laboratory / California Institue of Technology Eric Pottier, IETR, UMR CNRS 6164 Pascale Dubois-Fernandez, ONERA IGARSS’11

Overview • Definition of compact polarimetry mode • Calibration of a compact-pol system • Simulation of compact-pol data from full-pol raw data • Estimation of biomass with compact-pol data IGARSS’11

Issues

• Compact polarimetry

– 1 polarization on transmit – 2 polarizations on receive

• What is the best polarization on transmit?

• What are the best polarizations on receive?

• How do we analyze the data?

– Calibration – Faraday Rotation – Geophysical parameter estimation IGARSS’11

Background - Example with ALOS system Mode Swath Resolution Incidence angle 10m 8 ° ~ 60° HH HH/HV or VV/VH (dual-pol) Full polar (quad-pol) 70km 70km 30km 20m 30m 8 ° ~ 60° 8 ° ~ 30° • • • Single polarisation  Full polarisation  large swath and larger incidence angle range added characterisation Compact polarisation  full investigation of the dual-pol alternative IGARSS’11

Background - Compact Polarimetry 1/2 • π • π /4 mode: one transmission at /2 mode: one circular 45 ° and two coherent polarizations in reception (linear H & V, circular right & left,…) transmission and two coherent polarizations in reception (linear H & V, circular right & left,…)

k



k

   1 2  

S HH S VH S HV S VV

     1

j

   1 2  

S HH S VH

 

jS HV jS VV

  • Hybrid polarity : particular case of π /2 : one circular transmission and two coherent linear polarizations in reception (H&V) IGARSS’11

• Background - Compact Polarimetry 2/2  /4-mode potentials: reconstruction of the PolSAR information (1) – Iterative algorithm based on: • Reflection symmetry • Coherence between co-polarized channels •  /2-mode potentials: avoid Faraday rotation in transmission (2) – Transmit a circular polarized wave – Show results about the reconstruction of the PolSAR information from  /2 mode – Applications possible (3) : • Faraday rotation estimate • Soil moisture estimate • Classification using the conformity coefficient • Hybrid polarity potentials: decomposition of natural targets (4) –

m

d method based on Stokes parameters (1) (2) (3) (4) J-C. Souyris, P. Imbo, R. FjØrtoft, S. Mingot and J-S. Lee,

Compact Polarimetry Based on Symmetry Properties of Geophysical Media: The



/4 Mode

, IEEE Transactions on Geoscience and Remote Sensing, vol. 43, no. 3, March 2005.

P. C. Dubois-Fernandez, J-C. Souyris, S. Angelliaume and F. Garestier,

The Compact Polarimetry Alternative for Spaceborne SAR at Low Frequency

, IEEE Transactions on Geoscience and Remote Sensing, vol. 46, no. 10, October 2008.

M-L Truong-Loï, A.Freeman, P. C. Dubois-Fernandez and E. Pottier

, Estimation of Soil Moisture and Faraday Rotation from Bare Surfaces Using Compact Polarimetry

, IEEE Transactions on Geoscience and Remote Sensing, vol. 47, no. 11, Nov. 2009.

R. K. Raney, IGARSS’11 , IEEE Transactions on Geoscience and Remote Sensing, vol. 45, no. 11, November 2007.

Overview • Definition of compact polarimetry mode • Calibration of a compact-pol system • Simulation of compact-pol data from full-pol raw data • Estimation of biomass with compact-pol data IGARSS’11

Calibration – Full-pol system • Full-pol system calibration : 7 unknowns δ 1 , δ 2 , δ 3 , δ 4 , Ω,

f

1 ,

f

2

M



j



e D R SR D T



N M



A

 

e j

   1 d 1 d

f

1 2      cos  sin  sin cos      

S HH S HV S VH S VV

    cos  sin   sin cos       d 1 4 d

f

2 3   

N

• The S matrix can be recovered:

S

  1   1  1

R D MD R R T

 1  • Distorsions can be retrieved with measures over known targets: – Trihedral, dihedral, transponder, natural targets, etc.

A. Freeman et T. Ainsworth,

Calibration of longer wavelength polarimetric SARs

, Proceedings of EUSAR 2008, Friedrishafen, Allemagne, June 2008.

S. Quegan,

A Unified Algorithm for Phase and Cross-Talk Calibration of Polarimetric Data – Theory and Observations

, IEEE Transactions on Geoscience and Remote Sensing, vol. 32, no. 1, pp. 89-99, January 1994.

J. J. van Zyl,

Calibration of Polarimetric Radar Images Using Only Image Parameters and Trihedral Corner Reflector Responses

, IEEE Transactions on Geoscience and Remote Sensing, vol. 28, no. 3, pp. 337-348, May 1990.

IGARSS’11

Calibration – Compact-pol system • Compact polarimetric system:

M

 1 2

j



T

   1

j

  

N

1

R D M R

1  1 2

SR D T

   1

j

  • The transmission defects cannot be corrected a posteriori • System needs to be of high quality before transmission • With a high-quality transmission  4 unknowns d 1 , d 2 ,  ,

f

1

M

 1 2

j



e D R SR R

     1

j

  

N

IGARSS’11

Calibration – Compact-pol system

M



Ae j



e



j

 1 2  

S HH S



HH

d 2  cos cos   d 1  

f

1 sin sin      

jS VV jS VV

  sin d 2  sin   d 1  cos

f

1   cos     

Ae j

  

S S HV HV

   

j

d

j

2   d 1

f

1      • Compact polarisation – 3 reference targets are necessary • Dihedral @ 0° • Dihedral @ 45° • Trihedral • Full polarisation – More unknowns – But less targets are required – Natural targets can be used – Acquisition of both HV and VH d 1  2

j

   

M D

0

RV M D

0

RH M D RV M D RH



M D RH M D RV M D

0

RV M D

0

RH

   

f

1 

M T RH M T RV

2

j



M D RH M D RV

  

j

2 ln    

M D

0

RH M D

0

RV A T A D

 

j M D

0

RV M D

0

RH M D RV M D RH

 2

j

   1    d 2 

f

1 2

M D

0

RH M D

0

RV

 d 1 *

f

1 

jf

1 IGARSS’11

Overview • Definition of compact polarimetry mode • Calibration of a compact-pol system • Simulation of compact-pol data from full-pol raw data • Estimation of biomass with compact-pol data IGARSS’11

Simulated compact polarimetric data • Simulation of CP data is necessary • How do we proceed?

– Two options: • From raw data • From processed data • Comparison between the two approaches

Example of raw data, range spectra HH {R;G;B}={HH;HV;VV}, SETHI data, L-band, Garons

IGARSS’11

S HH raw

Processing (corrections, antenna beam, etc.)

Process 1

Building compact polarimetric data

Process 2

S HV ra w M RH



S HH



jS HV S HH ra w

Processing (corrections, antenna beam, etc.)

S HV ra w

Hilbert transform 

jS HV ra w S HH pro S HV pro k RH ra w



S HH ra w



A

_

HV j A

_

HH S HV ra w

Processing (corrections, antenna beam, etc.) Calibration :

k RH p ro



A

_

HH

 

S HH S HH p ro j A

_

HV A

_

HH S HV p ro

  Calibration:

k RH



A

_

HH



k RH ra w

Raw data Processed data M RHpro M RH IGARSS’11

Building CP data - Process 1 / Process 2

Image of CP data from FP processed data, {R ;G ;B}={ M Rh_pro +M Rv_pro ;M Rh_pro ;M Rv_pro } Image of CP data from FP raw data, {R ;G;B}={ M Rh +M Rv ;M Rh ;M Rv }

0 IGARSS’11

Coherence between both images

1

Compact-pol - Process 2 / Process 2

FP data {R;G;B}={<|VV|²>;<|HV|²>;<|HH|²>} FP reconstructed {R;G;B}={<|VV|²>;<|HV|²>;<|HH|²>}

IGARSS’11

Overview • Definition of compact polarimetry mode • Calibration of a compact-pol system • Simulation of compact-pol data from full-pol raw data • Estimation of biomass with compact-pol data IGARSS’11

Backscattering coefficients and biomass – RAMSES P band data over Nezer forest (HV)

(HV) (RH) (RR)

IGARSS’11

Polarization HV HV RR RH

Biomass estimate – Nezer forest

RMS error (tons/ha) quadratic regression

5.8

6.2

6.6

12.2

RMS error (tons/ha) exponential regression

5.7

6.5

6.6

12.8

RMS error = 2.6 tons/ha (HV vs HV) IGARSS’11

Biomass map – Nezer forest 120 tons/ha

B HV

 205.8

e

0.1274



HV B HV

 178.01

e

0.1465



HV

IGARSS’11

B RR

 53.142

e

0.1626



RR

0

Biomass map – Nezer forest 120 tons/ha

Measured biomass B HV

IGARSS’11

B HV B RR

0

Biomass estimate with HV regression Using the HV regression as a reference, computation of the biomass with HV backscattering coefficient RMS error=20.1 tons/ha Bias=19.5 tons/ha IGARSS’11

Summary: systems implications • Compact-pol allows – To acquire larger swath (versus FP) – To access wider incidence angle range (versus FP) – To avoid Faraday rotation in transmission (versus DP) • Calibration – A solution with 3 external targets • Raw data – Equivalence between CP from FP raw data and from FP processed data • Biomass estimate – FP: RMS error for HV: 5.8 tons/ha – CP: RMS error for HV reconstructed: 6.3 tons/ha – CP: RMS error for RR: 6.6 tons/ha IGARSS’11

Thank you for your attention

IGARSS’11