Comparison of different methods of supervised classification with a special view on their effectiveness in geological studies: A case study in the Skavagh, eastern Azarbaijan

The Skavagh study area is located in the western of the Alborz-Azarbaijan folded belt. The dominant rocks in this area are volcanic and volcanic-sedimentary rocks with the Eocene-Oligocene age which have covered most of the area. As well, the late Oligocene intrusive bodies with intermediate composition have cut the rocky units of the area. Generally, metal mineralization in the Skavagh area is copper and manganese which has occurred in the Eocene volcanic and volcanic-sediments rocks. In this research, satellite images of Sentinel 2 and Landsat 8 have been used to perform different methods of supervised classification and prepare a geological map. In this study, after performing the necessary pre-processing, the supervised classification of the images was carried out in different ways, and finally, by conducting field studies, the best method was selected based on the geological evidence. Finally, after calculating the error matrix, the accuracy of the data, Kappa coefficient and by comparing with the index samples and sampling and the thin sections of the geology, the method was validated. After sampling and statistical analysis, it was determined that the spectral angle mapping method showed the best match with the geological units of the region.