Feature extraction for cloud analysis using satellite imagery data

Atmospheric products derived from the moderate resolution imaging spectroradiometer (MODIS) instrument are widely accepted as the state-of-the-art by the meteorological community. These products are more useful to the meteorological community than the traditional cloud labeling results. However, these products are not available at regular temporal frequency over one specific region. Algorithms that create MODIS products cannot simply be applied to other satellite data. Thus, an innovative method is needed to estimate MODIS-like products using geostationary satellite imagery, which has a higher temporal frequency. This paper presents a canonical coordinate decomposition (CCD)-based method to estimate MODIS channels using imagery from the geostationary satellite Meteosat 8. The estimated data can subsequently be used to arrive at cloud phase products.