Texture models for high-resolution ocean microwave imagery

Computer simulations of ocean passive microwave images and texture features are developed in the context with high-resolution radiometric observations. The technique is based on statistical and deterministic characterizations of spatial-variable fields (image scenes) containing ensembles of different radio-hydrophysical factors - surface roughness, foam, whitecap, spray, and bubbles. The corresponding microwave contributions are calculated using available hydrodynamic-electromagnetic models. In particular, we employ the well known resonance model for simulations of low-contrast signatures (∼2-3 K) due to ocean surface roughness and macroscopic dielectric models for simulations of high-contrast signatures (∼10-20 K) induced by foam/whitecap. Spatial and statistical combinations of these factors including the generation of fractal geometrical sets allow us to model a novel class of ocean microwave textures and distinct radiometric features. Some of them have been observed and analyzed during remote sensing experiments conducted by NOAA/ETL. Examples of the texture modeling and synthesis are presented and discussed in the paper.