Study on geographic ontology based on object-oriented remote sensing analysis

In the preprocessing course of spatial data, different departments always have diverse naming methods when describing the same geographical entity, due to different backgrounds and views of angle. There is also great difference among the feature sets which are used to describe concepts of geo-ontology, making it difficult to conduct semantic interoperation based on the theory of concepts reasoning in the information science. Consequently, this paper takes green land system for example and presents a reasoning method of geo-ontology based on object-oriented remote sensing analysis. We firstly establish an image hierarchical network system by using the object-oriented multi-scale segmentation technology. Then, the mapping from domain ontologies to image objects is realized by the maximum area method. Finally, through analyzing the features of image objects, the reasoning principles are built up, realizing the semantic interoperation between concepts of ontologies and image objects.