Polarization filtering of SAR data

摘要

In this paper, a systematic theoretical analysis of polarization filtering for a bistatic system is developed for two applications: maximization of signal to noise ratio and discrimination between two target types. The described method finds the optimum receive antenna polarization analytically but relies on a numerical approach to find the optimum transmit antenna polarization. The method uses the Stokes matrix representation, and therefore applies to analyze the partially polarized scattered field from extended targets. Examples of the technique are presented for the NASA CV990 polarimetric L-band radar. Image enhancement filters maximizing the signal-to-noise ratio are developed for different noise characteristics and different target types (urban and forest). A filter is also developed to maximize the power ratio between urban and natural targets. Results show that the filter maximizing the contrast between urban and forest area is essentially the same as the one maximizing the contrast between urban and ocean areas.