Sea Wind extraction from SAR images

In this paper we give a review on the various techniques available at this moment for the extraction of the sea surface high-resolution wind field from Synthetic Aperture Radar (SAR) images of the ocean. Most of them rely on external information to univocally determine the wind vector. We present a methodology which does not need any external information to univocally retrieve the wind vector, provided longitudinal wind-related features are present on the image. It is based on the two dimensional Continuous Wavelet Transform (CWT2), which is used to decompose the image on elementary components in the spatial scales domain. The components are classified according to their energy content in the scales domain, so that an explicit hierarchy of the more energetic spatial features can be obtained. The reconstruction of the image amplitude is thus performed by selecting the appropriate components. One application of the method is also shown.