Detection of land-cover and land-use change in the Hohe Tauern National Park (Austria) using an object-based classification approach: first results

摘要

Land-cover and land-use (LCLU) change reflects environmental as well as social changes. High-mountain environments such as the Hohe Tauern Range (Austria) and its National Park are very sensitive to these changes. Remotely sensed data are ideally suited to detect and analyze land-cover and land-use changes on different spatial and temporal scales. In this study a Landsat Thematic Mapper (TM) time-series from 1988 to 2003, aerial photographs and auxiliary GIS data were used detect land-cover and land-use changes in the National Park and its surroundings. The objective of the study was to develop and test a transferable method for change detection in comparable high-mountain environments. After applying a series of pre-processing steps, object-based image analysis, on the basis of Definiens Developer, was used to extracted land-cover and land-use information from the Landsat images. To generate meaningful image objects, a multiresolutions segmentation algorithm was combined with image object merging. The resultant image objects were then classified by using a hierarchically organized classification method. In a further step the resultant outputs of the land-cover and land-use classifications build the basis for the post-classification change detection. Moreover the advantages and disadvantages of the developed method, the first results as well as future adaptations of the method and applications are discussed in this paper.