WIND DIRECTION EXTRACTION FROM SAR IMAGES USING NSCT TRANSFORM

Deriving the wind surface direction is one of the key steps to retrieve other features of wind from spaceborne Synthetic Aperture Radar (SAR). In this study first a preprocessing step in applied on the SAR image, to make the image ready for processing phase. NSCT transform is applied on the image for detecting wind direction on the next step. For wind direction estimation NSCT sub-bands that contain higher energy level, have been selected. Chosen sub-bands are clustered into different parts according to amount of energy in each pixel. The estimated wind direction has 180 degrees ambiguity, Which this problem is resolved by using the available models for wind direction estimation of SAR images. Two kinds of SAR images (1.COSMO-SkyMed (CSK) of the Italian Space Agency and 2.Sentinel-1A, VV polarization of the Strait of Bonifacio) have been analyzed and the results are compared with the model and scatterometer directions and the comparison indicated a good score of success.