Cost-Effective Raster Image Processing for Geoecological Analysis using ISOCLUST Classifier: a Case Study of Estonian Landscapes
Résumé
Current presentation demonstrates environmental analysis of the landscapes in Estonia, Eastern Europe. Methods include the use of Arc GIS 10.0 and IDRISI GIS Andes 15.0 for image processing. Research aim is o detect land cover changes using method of image classification 'ISOCLUST'. The raster processing GIS approach and classification was applied towards Landsat TM two images. The ISOCLUST is an unsupervised classification method in IDRISI GIS performs the most of the image processing workflow in semi-automatically regime. The study also reports photos of the Estonian landscapes. Results include 16 land cover types typical for the study area classified and visualized on the images. In 2006 the urban area became larger than in 1992 (land cover class "3" on the histogram. This can be explained by various reasons. Changes in land cover types in selected Estonian landscapes are shown on the statistical histograms on 1992 and 2006.
Domaines
Sciences de l'environnement Biodiversité et Ecologie Synthèse d'image et réalité virtuelle [cs.GR] Interface homme-machine [cs.HC] Ordinateur et société [cs.CY] Apprentissage [cs.LG] Modélisation et simulation Traitement des images [eess.IV] Environnement et Société Milieux et Changements globaux Ingénierie de l'environnement Informatique [cs] Vision par ordinateur et reconnaissance de formes [cs.CV]
Origine : Fichiers produits par l'(les) auteur(s)