Creating excess water inundation maps by sub-pixel classification of medium resolution satellite images

Abstract

Excess water frequency factor, which indicates the number of inunda-tions in the area under study within a certain period of time, is the most dynamic variable among the parameters applied in the complex meth-odology of excess water hazard mapping. Creating excess water inun-dation maps, representing the situation in the most realistic way, was hitherto a critical moment in excess water hazard mapping. Instead of field measurements, since the database of Landsat satellite images became accessible in 2009, it is possible to process satellite images taken from the year 1985, with using new, non-traditional methods different from the pixel-based classification. These methods are mainly sub-pixel based classifications and they are applied principally on images taken in periods of extended excess water inundation, under clear weather conditions. In our research project, medium-scale map-ping was supported principally by hand-held or mounted multispectral (the bands of visible and infrared light) digital aerial photography. The photo-taking process, depending on the actual meteorological condi-tions, can be flexibly accomplished in the most extended inundation period, thus it is possible to create excess water maps at the scale of 1:10000.