期刊名称：GI_FORUM - Journal for Geographic Information Science
电子版ISSN：2308-1708
出版年度：2013
页码：187-196
DOI：10.1553/giscience2013s187
出版社：ÖAW Verlag, Wien
摘要：Automatic feature extraction from satellite imagery is cost effective and fast. An essentialissue in this context is the degree of accuracy for thematic correctness obtainable throughcommon pixel-based and object-oriented classification algorithms. By applying two classificationalgorithms to Landsat 5 TM imagery for the extraction of different morphologicalriver features the thematic correctness of the resulting raster images and the separability ofthe river features is evaluated.River features of meandering rivers evolve through dynamic avulsion, erosion and depositionprocesses. Although many studies focus on the analysis of these river environments,diverse methods of GIS and remote sensing based river feature classification methodshave not been evaluated and assessed yet. In the literature several techniques to monitorspatio-temporal changes such as lateral river channel migration are already mentioned butthe tendency there is to identify the changes by examining time spans rather than a point intime. Besides that the semiautomatic river feature methods described in related studiesmainly focus on the identification of a river channel itself and do not consider additionalfeatures such as oxbows, scars, relic channels, etc. that in fact are significant characters inriverine environments. Therefore, this paper evaluates the application of a supervisedclassification using ENVI’s Support Vector Machine and an object based classificationusing the ArcGIS extension Feature Analyst to extract river features from Landsat 5 TMimages including ancillary data files. Furthermore, the results of the classification methodsare evaluated with regard to thematic correctness and separability of the various classifiedriver features using accuracy assessment as presented in the specialist literature. Finally thelong-time changes in the riverine environments are traced by interpreting the distribution ofthe classified river features. Accordingly, the approach of this work contributes to on-goingresearch concerning semiautomatic or automatic river feature extraction.
