You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.
 Earthnet Home  
Sessions and Session Summaries
First Annoucement
Scientific Committee
ESA Specific Links
Organising Committee
Round Table Discussion Questions
Conference Photos
List of Participants
All papers


What can be learnt from a half-polarimetric SAR?

Dr Jean-Claude Souyris (1), Mr Patrick Imbo(1) , Dr. Jong-Sen Lee(2) , and Ms Sandra Mingot(1)

(1) CNES, BPI 811, 18 Avenue Edouard Belin, 31401 Toulouse Cedex 4, France
(2) NRL, 4555 Overlook Ave. SW, Washington DC 20375-5351, United States


The overall objective of this study has been to assess various designs of partial polarimetry (pp) architectures that could be implemented on space segments at a reduced cost, and compliant with micro-satellite platforms (»100 to 120 kg class). The performance assessment concerns both the level of information preserved in comparison with full polarimetry (fp), and the concurrent space segment complexity, in terms of PRF, processed swath, and down link features.

In previous studies, two pp options were investigated. The first one is limited to a unique linear polarization in transmission (e.g. horizontal), while receiving on both co- and cross-polarized channels (h and v). This option is based on a simple design, thanks to the polarisation uniqueness in transmission. However, it led to poor classification performances [1]. The second option considers two interleaved polarizations in transmission (h and v), the reception being restricted to co-polarized channels (i.e. hh and vv). The polarimetric information is much better preserved than in the previous case [1], [2], and the power budget is more comfortable (hv term which is now left out is usually 7 to 10 dB below hh and vv). However, the main drawback is linked to the need for transmitting two polarizations : in comparison with a standard SAR, the PRF has to be doubled, and the swath halved, in order to maintain immunity from range ambiguities.

The framework of this study is consequently to restrict pp architectures to the polarization uniqueness in transmission, combined with the polarization duality in reception. Such a configuration justifies the denomination of 'half polarimetry'. In this context, in order to overcome the drawbacks of the (hh+hv) architecture (critical loss of information), we focus on a configuration where the two linear receiving polarizations are oriented at p/4 of each side of the unique transmitted linear polarization. A procedure to estimate a full polarimetric (fp) information over extended targets from this design (hereafter called the p/4 mode) is proposed, when the nullity of complex correlation between co- and x-polarized terms is assumed. The behavior of point targets is also considered.

As already mentioned, the polarization uniqueness of the p/4 mode in transmission permits to maintain the PRF at the same value as for a standard SAR. For a given swath, the down-link features are reduced by a factor of 2, in comparison with fp. The power balance of the p/4 mode is far better than for (hh+hv), for which the two receiving channels are affected by critically different Signal to Noise Ratio (SNR), X being typically 7 to 10 dB below H and V. However, a drawback of the p/4 mode is related to the non parallelism of transmitted and received fields, leading to a loss of 3 dB on the power link. At the image level, the Noise Equivalent s0 (NEs0) increases by 3 dB, transforming a typical value of -25 dBm²/m² to -22 dBm²/m². Further studies must be conducted to evaluate the real impact of this loss.

At the time of the presentation, the assessment of the p/4 mode will be discussed for its crop classification capability using L and C band images acquired over the Flevoland area by JPL AIRSAR. In order to assess the performances of the various pp architectures (including the p/4 mode) in comparison with fp, the response of p/4 mode is simulated using the fully polarimetric SAR data of Flevoland. Quantitative comparisons of classification accuracies between fp, pp and the p/4 mode are evaluated based on the maximum likelihood Wishart classification. In our initial study, the p/4 classification accuracy is comparable to that of the fp and the (hh+vv) mode, and is much better than the (hh+hv) mode.


[1] : J.S. Lee et al. "Quantitative comparison of classification capability : fully-polarimetric versus partially polarimetric SAR", IEEE Trans. on Geoscience and Remote Sensing, vol. 39, n0. 11, pp. 2343-2351, November 2001.

[2] : P. Imbo, J.C. Souyris, "Assessment of partial polarimetry versus full polarimetry architectures for target analysis", proceedings EUSAR 2000.


Workshop presentation


  Higher level                 Last modified: