Detecting vegetation changes in a wetland area in Northern Germany using earth observation and geodata

摘要

Monitoring land use and landscape dynamics in conservation areas is important to understand and influence nature and restoration processes. Earth observation data can help to detect changes automatically in extensive areas. In a wetland area in Northern Germany different change detection methods have been tested to detect wetland restoration processes, especially succession of wetland and moorland vegetation over 11 years. Therefore a change detection method based on a selective principal component analysis followed by a fuzzy membership function introduced by Weiers et al. (2001). was tested with dual date Landsat TM/ETM+ images. As comparison vegetation maps and Colour-infrared (CIR)-aerial photographs were analysed. The main objectives were to find out (1) if changes, especially vegetation changes, can be detected on the study area by the method as described by Weiers et al. (2001), (2) which changes can be detected and (3) which is the best method on the study area, respectively: the Landsat change detection method, the analysis of vegetation maps or the interpretation of CIR-aerial photographs. For detecting vegetation changes the most detailed information were achieved by interpreting CiR-aerial photographs, while the Landsat change detection method turned out to be more suitable for detecting changes of wetness.