Evaluating land cover types from Landsat TM using SAGA GIS for vegetation mapping based on ISODATA and K-means clustering

The paper presents the cartographic processing of Landsat TM image by two unsupervised classification methods SAGA GIS: ISODATA and K-means clustering. approaches were tested compared for land cover type mapping. Vegetation areas detected separated from other types in study area southwestern Iceland. number clusters was set to ten classes. satellite GIS achieved using Imagery Classification tools Geoprocessing menu GIS. Unsupervised performed effectively unlabeled pixels machine learning Following an iterative approach clustering, grouped each step algorithm reassigned as centroids. contributes technical development application cartography demonstrating effectiveness remote sensing data applied vegetation environmental