214 Information Extraction of Coastal Landscape and its Response to Dam Disturbance in Yellow River Delta Based on Remote Sensing Images

In this paper, the landscape cover, the coastline and artificial dams were extracted from Landsat 5 TM and SPOT5 images by the combination of supervised classification based on Support Vector Machine (SVM) and visual interpretation. And the influence of artificial dams on landscape structure and landscape pattern was analyzed by landscape pattern index and gradient analysis of NDVI. The results suggest that, the accuracy of landscape cover extraction met the requirements of change analysis of the landscape pattern. The artificial dams effectively prevented the backward erosion of inland area. In addition to the estuary area, the area of the tidal flat landscape was reduced a lot in the region without dams. Spatial and temporal analysis of landscape pattern indices showed that dam disturbance made the landscape structure changed largely and the degree of landscape fragmentation increased significantly. The change rates of mean area of patches (AREA_MN) and splitting index (SPLIT) were 1’0.79 and 6.98, respectively, which were higher than that of the region without dams. The trend of the gradient changes of NDVI in the region with dams on the distance from the coastline was nonlinear, while linear in the region without dams.