Teodóra BATA¹, Gábor MEZŐSI² & Burghard C. MEYER^3
1University of Szeged, Department of Physical Geography and Geoinformatics, 6722, Egyetem Street 2-6, Szeged, Hungary, e-mail: batateodora@gmail.com
²University of Szeged, Department of Physical Geography and Geoinformatics, 6722, Egyetem Street 2-6, Szeged, Hungary, e-mail: mezosi@geo.u-szeged.hu
³Geographical Institute, University of Leipzig, Johannisallee 19a, 04103 Leipzig, Germany, e-mail: burghard.meyer@olanis.de

LANDSCAPE UNITS FOR HUNGARY USING MULTIRESOLUTION SEGMENTATION OF GEO-DATA AND FUZZY ANALYSIS

Full text

The result of the study is an objective delimitation of landscape units by combining mathematical-statistical multi-objective segmentation methods and digital maps using geo-statistics and a variable number of shaping factors. The starting point is the important scientific replication problem of traditional or subjective landscape units developed during recent decades (e.g., for Germany and Hungary) via the bordering of landscapes into macro-, meso- and micro-regions. Subjective and objective segmentation methods were compared using quality indicators. The study is based on fuzzy indicator maps of the shaping factors of the soil, lithology, potential natural vegetation, soil water management, hemeroby and slope. A multiresolution segmentation analysis was applied to Hungary to delimit the borders of the meso-scalic landscape units. The complex results were analysed and compared with two traditional Hungarian landscape delimitations and with the results of the European-wide Landscape Classification (LANMAP). The results of the regional comparative analysis first show that the multiresolution segmentation method is applicable at the micro-region scale level and additionally demonstrate that the segmented landscape units offer a landscape unit system for Hungary that is an alternative to the traditionally defined landscape units, which are scientifically validated to only a minor extent and are related to few past applications in territorial or landscape planning. The applied multiresolution segmentation methodology combined with fuzzy set methods and based on multiple layers of geodata is an acceptable and useful method to delimitate landscape units at various hierarchical levels.