“…Fulajtár (2001) noted that additional ancillary data is needed for the classification of erosion patterns with required accuracy, and other authors have also recommended a combination of automatic classification approach and visual interpretation (Báčová & Krása, 2016;Pilesjoe, 1992;Šarapatka & Netopil, 2010;Smetanová, 2009). The application of the fuzzy classification (Meléndez-Pastor, Pedreño, Lucas, & Zorpas, 2017), the spectral mixture (sub-pixel) analysis (Haboudane, Bonn, Royer, Sommer, & Mehl, 2002;Rabah & Farah, 2016;Schmid et al, 2016) or the object classification ("spatio-contextual" image classification) (Mayr, Rutzinger, Bremer, & Geitner, 2016;Nobrega et al, 2006;Wang, Huang, Du, Hu, & Han, 2013) represents another solution to overcome the above mentioned difficulties related to the pixelbased methods (Li et al, 2014). Simultaneously, using higher resolution data in the spectral domain (hyperspectral data) promises an increase in the accuracy of the classification of erosion-degraded soils (Chabrillat et al, 2003(Chabrillat et al, , 2014Haubrock, Chabrillat, & Kaufmann, 2004Hill et al, 1994;Schmid et al, 2016;Žížala et al, 2017).…”