1. MODEL-BASED POLARIMETRIC SAR CALIBRATION METHOD USING FOREST AND SURFACE SCATTERING TARGETS Masanobu Shimada Earth Observation Research Center (EORC) Japan Aerospace Exploration Agency (JAXA), Sengen 2-1-1, Tsukuba, Ibaraki, Japan, 305-8505, Voice 81-50-3362-4489, Fax: 81-29-868-2961, shimada.masanobu@jaxa.jp July 28, 2011 IGARSS2011
2. POLCAL history Van Zyl (1990): radar reciprocity, scattering reciprocity(HV=VH),reflection symmetry, and CR Quegan (1994): scattering reciprocity, no-radar reciprocity, reflection symmetry, and CR Ainsworth (2006), Kimura(2008):reflection symmetry and use of urban areas Motivation of research: Rain forest : distributed target (diffuse scatterer) could be usable for the POLCAL reflection symmetry ? , scattering reciprocity? >New trial of POLCAL using the natural target and the surface scatterer (CR)
3. Covariance model for forest Forest data covariance is expressed by double bounce, surface scattering and volume scattering. (S. R. Cloud, and E. Pottier,”A Review of Target Decomposition Theorems in Radar Polarimetry,” IEEE Trans. on Geoscience and Remote Sensing, Vol. 34, No. 2, March 1996, pp. 498-518) Asummption:
6. L-band backscattering from the forest consists of the volume scattering and double bounce (polarization dependent). (reflection symmetry: no-correlation between like and cross , i.e., <HH*HV*>=0), large data available.
21. Ignorance of the surface scattering Condition: Freeman-Durden ignores HV of surface scattering This models ignores the surface scattering. Do these ignorance affect the parameter estimation? These ignorance do not affect the parameter estimation.
35. Fig. 19. Incidence angle dependence of the forest parameters estimated from the proposed PolCAL method. Within a 3-degree incidence angle range, two components, the volume scattering component, and the double-bounce component, are almost independent of the incidence angle. The amplitude ratio of the double-bounce component in HH to that in VV is nearly equal to 2.1.
36. Comparison of the interferometric coherences in HH and VV Fig. 22. Comparative display of repeat-pass SAR interferometric coherences of HH-HH in a) and VV-VV in b). The master and slave images were acquired on Oct. 20, 2006, and Sept. 6, 2006, respectively as 46 days separated in time and 193m separated in space (e.g., perpendicular baseline). Slightly dark stripes running vertically seem to be interferometric decorrelation due to ionospheric disturbances [29].
37.
38. 26 datasets for the Amazon test sites showed the good stability of the PolCAL parameters.
44. Fig. 15. Temporal variation of the volume scattering component and the double-bounce scattering components of VV and HH. Double-bounce component of HH is nine times larger than that of VV. This calculation is obtained by Method-1”’.