Remote Sensing Image Land Cover Classification And Change Detection Based On Geo-ontology

The method research of remote sensing image classification and change detection is always an important topic in the field of remote sensing.With the development and research of remote sensing technology,classification and change detection technique on the respect such as the extraction of effective feature information,monitoring the state of the union,database updating etc...provide a strong support,and gradually become the key technology of social production and economic development.With the development of remote sensing image,The geographic features lack of conceptualization and formal expression in classification and change detection,different features has different configuration,which often leads to the difference of classification and change detection results,how to implement conceptualization and formal expression of different features,how to implement geographic features to build model objectively to organize feature information knowledge effectively and express formally,which has become a difficult and hot task on the application study of change detection.In view of this problem,the paper mainly focuses the method research of classification and change detection,uses high resolution image as research object,puts forward construction of geographic ontology semantic model method based on feature information knowledge,combines the advantage of multi-source data,gets feature information knowledge from multi-source,introduces the theory of geographic ontology,implement geographic features to build model objectively,to organize feature information knowledge effectively and express formally,implement conceptualization?formal expression of different features,builds feature knowledge database,conducts geographic ontology model conducts the method research of classification and change detection based on it,tests remote sensing image classification and change detection experiment and pseudo-change recognition experiment,the validity and practicability of the model and method are proved.In some sense,the method provides an effective way for the high resolution image classification and change detection.