Cost-Effective Raster Image Processing for Geoecological Analysis using ISOCLUST Classifier: A Case Study of Estonian Landscapes
Modern Problems of Geoecology and Landscapes Studies. Proceedings of 5th International Research Conference. Belarus State University (BSU). Ed.: A. N. Vitchenko et al. Minsk, Belarus: BSU Press, 2014, pp. 74–76. ISBN: 978-985-476-629-4
3 Pages Posted: 10 Jan 2019
Date Written: October 14, 2014
The goal of the study is to perform comparative analysis of image processing methods, enabled by means of IDRISI GIS software. The purpose has two aims. First, a spatial analysis of land cover types in the coastal landscapes of western Estonia, Pärnu area, at two various temporal dates. Second, an overview of the technical methods of IDRISI GIS enabling to evaluate image processing. The main research method consists in classification of satellite images for resulting geoecological landscapes mapping. The ISOCLUST classification enabled to create multispatiotemporal thematic maps of Pärnu area. The research method is based on the spatio-temporal analysis of the geospatial data, performed by means of GIS tools and remote sensing data.
The analysis of spectral reflections strongly depends on the local inner features of the surface: texture, structure, colour, etc. Information on spectral signatures in received by the satellite sensors and recorded on the Landsat TM. This information is used for the image classification. Using individual characteristics of objects, derived from the multi-spectral Landsat TM bands, the information from the image was extracted, analysed and used for the classification. Since there are 16 classes typical for the land cover types in the study areas, according to CORINE, the predefined number of classes for ISOCLUST classifier was 16.
Using individual characteristics of objects, derived from the multi-spectral Landsat TM bands, information from the images was extracted, analysed and used for land classification. Comparing both images, changes in landscape patterns on 1992 and 2006 can be noticed. It is shown that between 1992 and 2006 there are changes in land cover types in selected Estonian landscapes. In 2006 the urban area is larger than in 1992. This can be explained by various reasons, e.g. intensive suburbanization in the tourist zone of Baltic Sea, and increase of summer cottage («dacha») construction on the coasts which increased settled areas.
Keywords: land cover change, environmental analysis, climate change, Estonia, landscapes, mapping
JEL Classification: Y92, Y10, Q01, Q24, Q51, Q54, Q57
Suggested Citation: Suggested Citation