SAR-ocean wave inversion using image cross spectra

Abstract
Image cross-spectra obtained by combining pairs of single look SAR images are utilized in an inversion scheme for extracting the underlying ocean wave spectrum. The reasons for proposing the use of image cross-spectra instead of standard multilook spectra are twofold. First, the image cross-spectra are shown to significantly reduce the speckle noise level while preserving the spectral shape. Second, the image cross-spectra provide information about the wave propagation direction. A closed form expression of the ocean-to-SAR spectral transform and its inversion is formulated for the cross-spectral case. The inversion method is based on computing the full derivative of the nonlinear ocean-to-SAR image cross-spectral transform. The SAR image cross-spectra are shown to yield a SNR typically 16-25 dB higher than the corresponding standard 3 look image spectra. The noise level is typically reduced with a factor of 20 compared to the 3 look spectra and a factor of 60 compared to the single look spectra. The inherent information about the wave propagation direction is utilized in the inversion scheme to resolve the 180-degree ambiguity without using any external information. The method is demonstrated on airborne C-band SAR data.<>

This publication has 15 references indexed in Scilit: