Classification of Urban Land Use Based on Graph Theory and Geographic Information System

In recent years, the time dimension of geographic objects has been introduced to the geographic information system (GIS), marking a great progress in geographic information science (GIScience). The mining and knowledge discovery of spatial pattern and spatiotemporal evolution of geographical elements and phenomena have become the hotspot of GIScience research. Based on graph theory, abstract definitions were given to geographical simple objects and composite objects, involving spatial relationships, morphological features and semantic features. These features, attributes and relationships were integrated into a novel relational attribute neighborhood graph (RANG) through the design of a series of algorithms. The RANG was successfully applied to the automatic classification of urban land use, providing a desirable tool for land use classification in the context of rapid urbanization.