Edge pixel classification at subpixel accuracy through resolution enhancement and deconvolution processing, in Landsat-TM imagery

A great number of efforts have been made to improve the classification results of low-resolution satellite imagery. The main interest up to now was focused on how much is present at each mixed pixel of every class. The main aim of this paper is not only to determine how much, but also to tackle the problem of where the classes are located within the pixel in a graphic way. The methodology proposed is based on a spatial resolution increase by means of a combination of interpolation and deconvolution processes. The test bed consisted of images acquired by the Thematic Mapper (TM) sensor on board the Landsat V satellite