A Research On Spatial Data Mining Of Digital Elevation Models

Digital Elevation Model (DEM) is widely known as the kernel of National Foundation Geographical Information Database, as well as the basis of digital terrain analysis in GIS for it contains plentiful topographical, geomorphology and correlative geographical information and acknowledge. Nowadays, DEMs at four map-scale levels of have been constructed in China. Faced on the huge quantity information saved in the databases, we believe that a careful researching and mastering the theories and techniques of spatial data mining from DEMs have vital significance. Applying the basic theories of spatial data mining and geomorphology, taking the Loess Plateau of North Shaanxi Province, China as the research area, this dissertation explored the theories and systematic structures of geomorphologic data mining and acknowledge discovery from DEMs, ascertained the systematic structures of topographic factors and synthetic topographic acknowledge, as well as systemically summarized mining algorithms from DEMs. The author paid more attention also to analyze the principle and algorithm of flow length, curvature, relief, the earth's surface incision and gully density. In this research, the elementary constitutes and mining algorithms of regional character acknowledge in the Loess Plateau were probed. It is on ArcView GIS and DelphiT.O software, some extended functions for DEM spatial analysis and engineer assistant system of geomorphologic data mining were added. A careful statistic and analysis were done in exploring the terrain factor's spatial distribution in the Loess Plateau of North Shaanxi Province. The results of research proved that, the theories and techniques improved in this research are of theoretical significance and show a splendid application foreground.