The Earth’s surface is a fundamental variable that connects many parts of the global environment. Therefore, the existence and distribution of different land cover types must be understood in order to categorize and predict their potential impact on the environment. For that purpose, this paper applies the unsupervised automatic classification and use of the Iterative Self-Organizing DAT Analysis (ISODATA) method to detect land cover types, based on multispectral satellite image from Landsat-8 from 2018 for the valley of river Treska in Republic of Macedonia. The final product is a land cover classification map with four classes: Water bodies, Forests, Shrubland and Herbaceous vegetation.

The general aim of this paper is to understand and emphasize the importance of land cover allocation, which is important for maintaining the daily flow of living, without having to perform terrestrial measurements (which reduces the time and costs required for acquisition), using geospatial tools and remote sensing methods to categorize the land cover classes.