Three-channel Wiener filter for suppression of speckle noise in SAR images of ocean scenes

Extraction of wave-height information from synthetic aperture radar (SAR) images of ocean scenes is difficult due to the presence of speckle noise. It has been observed that the auto-correlation of the speckle noise is not negligible at non-zero lags; this violates the basic requirement of most noise suppression techniques. This paper describes an approach which uses three-channel Wiener filtering to suppress the speckle noise and obtain an image in which the ocean wave structure is greatly enhanced. The filter is designed to operate in a “gapped” prediction mode, and is applied sequentially to the rows and columns of the SAR image. The performance of the technique is demonstrated with data from an airborne SAR