ABSTRACT:In this work, an object-oriented image analysis methodological framework employing edge extraction and image processing techniques was proposed and applied on a LANDSAT-ETM+ image of Alevrada, Greece to derive lineament classes. For the design of the knowledge base, the input data layers to the lineament identification system were (1) an edge map from the EDISON edge extraction algorithm performed on band 5 of the LANDSAT-ETM+ image, (2) the geologic layers derived from the geologic map at the pre-processing stage, (3) the initial ETM+ image of the study area and its derived thematic products using remote sensing methods (such as NDVI, PCA and ISODATA unsupervised classification) for the discrimination of the land cover classes. Segmentation was performed based on the multi-scale hierarchical segmentation algorithm in eCognition for the extraction of primitive objects of the input data. Finally, intrinsic spectral and geometric attributes, texture, spatial context and association were determined for the designed object classes / sub-classes on each segmentation level, and fuzzy membership functions and Nearest 384 O. Mavrantza, D. Argialas Neighbor Classification were employed for the assignment of primitive objects into the desired thematic classes combining all participating levels of hierarchy. The output results of the system were classification maps at every hierarchy level as well as the final lineament map containing the geological lineaments of the study area (possible faults) and the nontectonic lineaments.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.