Geographic ontology driven hierarchical semantic of remote sensing image

Remote sensing study focuses on not only information of image its own, but also geographic analysis based on these information. Unlike normal image, remote sensing image holds several special characteristics, which make impossible to create hierarchical semantic of remote sensing image only applying classic achievements on image hierarchical semantic. In this paper, we firstly indicate two characteristics of remote sensing image hierarchical semantic and provide processing from feature level to semantic level. Then depend on theories of geographic ontology, we design the framework of remote sensing image hierarchical semantic and test our framework through typical experiments. Overall, it proves that, compares with normal image hierarchical semantic, our remote sensing image hierarchical semantic owns ability to handle with the issue of both spatial, attribute, temporal semantic and logic reasoning. Result reveals that our idea holds much better in completeness and robustness driven by geographic ontology.