In this paper, we are investigating the optimization of frequencies and polarizations from the viewpoint of target detection in SAR images. Frequency and polarization diversity provided by multifrequency polarimetric SAR opens new ways for more robust detection of ground stationary man-made targets in complex clutter. Due to the scene inhomogeneity, the scattering characteristics of the area within each SAR resolution cell may be nonuniform. Each nonuniform pixel have multiple elementary scattering centers belonging to several classes. Radar returns for each pixel are contributions of the collective interactions of electromagnetric waves scattered by all scattering centers within the pixel. Multifrequency polarimetric measurements for each pixel makes it possible to decorrelate signals from different classes of scatterers within the resolution cell. For the general target-plus-clutter versus clutter case, in this paper is established a general model for target detection in clutter using multi-look multifrequency polarimetric SAR measurements only for the pixel under test. Based upon this model, typical polarimetric covariance matrix parameters at several bands are used to illustrate the procedures to optimize frequencies and polarizations for target detection in clutter.

Date Published: 21 December 1994
PDF: 10 pages
Proc. SPIE 2316, SAR Data Processing for Remote Sensing, (21 December 1994); doi: 10.1117/12.197538

Published in SPIE Proceedings Vol. 2316:
SAR Data Processing for Remote Sensing
Giorgio Franceschetti, Editor(s)