Application of Different Methods of Decision Tree Algorithm for Mapping Rangeland Using Satellite Imagery (Case Study: Doviraj Catchment in Ilam Province)
Authors
Abstract
Using satellite imagery for the study of Earth’s resources is attended by many
researchers. In fact, the various phenomena have different spectral response in
electromagnetic radiation. One major application of satellite data is the classification of
land cover. In recent years, a number of classification algorithms have been developed for
classification of remote sensing data. One of the most notable is the decision tree. The aim
of this study was to compare three types of decision trees split algorithm for land cover
classification in Doviraj catchment in Ilam province, Iran. For this, propose, first, the
geometric and radiometric corrections were performed on the 2007 ETM+ data. Field data
as training sites were collected in the various classes of land use. The results of image
classification accuracy assessment showed that the Gini split classification. With kappa
value 89.98 and the entire accuracy 91.17% was significantly higher, then categorization of
branching and the branching ratio and Entropy with kappa values of 88.45 and 90.65 and
the entire accuracy of 86.21 and 86.15%, respectively.
