Polarimetric Palsar System Model Assessment and Calibration

Ridha Touzi(1), Shimada Masanobu(2), Stefan Nedelcu(1,1) and Masanobu Shimada(2)

(1) Canada Centre for Remote Sensing, 588 Booth St.,, Ottawa, Ontario, K1A 0Y7, Canada
(2) Japan Aerospace Exploration Agency, Sengen 2-1-1,, Tsukuba, Ibaraki, 305-8506, Japan


Like L-band JERS-1 SAR, ALOS PALSAR is affected by Faraday rotation. The eventual presence of Faraday rotation in addition to the uncertainty regarding the actual isolation of the H-V PALSAR antenna, lead to mixed conclusions (from the ALOS Cal-VAL team) regarding the actual isolation of the H-V PALSAR antenna. In the last Pol-In SAR’07 workshop, it was concluded that the antenna isolation is not better than -25 dB. Such a low isolation should make the system calibration more complicate in the presence of Faraday rotation, mainly for areas with significant topographic relief.

In this paper, ALOS-PALSAR data sets collected over various calibration sites are used for the assessment and validation of the polarimetric PALSAR system parameters. In each site, one or several corner reflectors have been deployed during the acquisition, and the data collected over the Amazonian forest, the CCRS-Ottawa calibration site, as well as the calibration sites of DLR (Germany) and the Chalmers university of technology in Sweden are investigated (DLR and the Chalmers university of technology are thanked for having kindly provided us with the data). First, the PALSAR system parameters are assessed under low Faraday rotation conditions, using the Amazonian forest data sets. The Freeman-Van Zyl calibration technique [1], which symmetrizes the system prior to the estimation of the distortion matrix elements, is adopted for this study. Van-Zyl’s technique for determination of system distortion matrix elements is reconsidered, and additional equations [2] are used to optimize the estimation of the calibration system unknowns even when the azimuthally symmetric reference target is of very low HV return in comparison with HH, VV, and the HH-VV cross-correlation. The new method is validated, and the accuracy of system parameter retrieval is discussed as a function of the azimuthally symmetric target used as a reference target. The method is then applied to two data sets collected in July 2006 over the Amazonian forest calibration site, with a corner reflected deployed for calibration purposes. It is shown that the system is highly isolated with an antenna H-V isolation better than -38 dB. The use of distortion matrix of lower isolation (-23 to -30 dB) computed at the presence of small but still significant Faraday rotation angle (2-3 degree) induces a significant return (-23-25 dB) at the cross-polarization of the corner reflector. Such significant return (-23-25 dB) might be interpreted as the antenna H-V cross-talk when the Faraday rotation is assumed to be zero, as done at the Pol-In SAR’07 and this explains the miss-leading conclusions regarding the low isolation of the PALSAR H-V antennas (lower than -25 dB [8]). The results obtained at the Amazonian forests were confirmed with the Sweden data at very low Faraday conditions (during night), and it is found that even a distortion matrix with very low cross talk (-35 to -37 dB) might introduce small errors in the Faraday rotation estimate. The best results are obtained with the assumption of zero antenna cross talk (diagonal antenna distortion matrix), and this should ease the JAXA preliminary calibration which would have to correct only for antenna gains incidence variations (zero cross talks).


[1] A. Freeman, J.J. Van Zyl, J.D. Klein, H.A. Zebker and Y. Shen, “Calibration of Stokes and scattering matrix format polarimetric SAR data”, ”, IEEE Trans. Geoscience Rem. Sens., Vol. 30pp. 531-539, May 1992. [2] Touzi R., C.E. Livingstone, J.R.C. Lafontaine, and T.I. Lukowski, “Consideration of antenna gain and phase patterns for calibration of polarimetric SAR data”, IEEE Trans. Geoscience Rem. Sens., Vol. 31, No. 6, pp. 1132 1145, Nov. 1993. [3] A. Freeman. Calibration of linearly polarized SAR data subject to Faraday rotation. IEEE Trans. Geoscience Rem. Sens., Vol. 42, No. 8, pp. 1617-1624, 2004. [4] R. Touzi and M. Shimada, “On the use of symmetric scattering target for calibration of ALOS” , Proc. of IGARSS’04, Alaska, USA, September 2004 [5] R. Touzi, M. Shimada, K. Papathanassiou and D. Corr, “On the use of point and distributed target scattering decomposition for calibration of ALOS” , Proc. of ASAR’05, Montreal, Canada, Nov. 2005 [7] Touzi, R., “Target scattering decomposition in terms of roll invariant target parameters”, IEEE Trans. Geoscience Rem. Sens., Vol. 45, No.1, pp. 73-84, Jan. 2007. [8] Pol InSAR’07, .



Symposium presentation


  Higher level                 Last modified: 07.05.06