Fuzzy Representation of Land Cover Classes Using Digital Geo-Data

The earth surface and human activities are interrelated in a complicated manner, and therefore it is somehow difficult to be monitored and managed. On this context it is required to develop effective directives of policies for environmental protection and management. In order to exploit context in imagery, a methodological approach employed data processing techniques was proposed in this work This approach was applied on a LANDSAT-TM satellite image and other digital geo-data (such as scanned geologic maps, topographic maps, etc) of a selected study area in Greece. The proposed scheme generally consisted of the following stages: (1) remote sensing methods and analysis for extracting the required thematic information to be used as input data and (2) fuzzy representation of the intrinsic land cover classes in the geodata through an object-oriented knowledge base design. The inherent land cover classes were represented by their intrinsic spectral and geometric attributes, texture, spatial context and association and fuzzy membership functions.