Reduction of Noise Interference from METEOSAT Water Vapor Image Data by Means of Fourier Transform and Frequency Domain Filtering

Abstract

Images provided by the water vapor channel data of meteorological satellites are suitable for the determination of wind vectors in the midtroposphere. Preprocessing the image data affects the quality and quantity of the derived wind vectors. Standard space domain filtering of image pixels with suitable algorithms gives rise to water vapor images which are usable for wind determination. However, if the superimposed noise shows obvious spatial frequency components, the image spectra make it much easier to attenuate those spatial frequency regions than is possible with the image itself. Then, interactive spatial frequency domain filtering after a Fourier transform of the original images allows better results for less effort. A space domain filter algorithm to tackle such a problem is, of course, conceivable. However, on the interactive system we use, constructing a space domain filter is much more difficult than constructing a spatial frequency filter.