The Construction And Application Of The Eco-geographical Regions Knowledge Database

With the continuous development of artificial intelligence,knowledge base and expert system have been gradually introduced into the field of remote sensing,especially for the identification and classification of remote sensing targets.The change detection of land cover was based on the remote sensing image at present.However,there are many errors and uncertainties because the images just reflect the instantaneous state of the surface.There are a lot of the phenomenon of different objects which have the same spectrum and the same objects have different spectrum especially when facing the global complex classes.Also due to the seasonal phase,there are a large number of pseudo changes in the process of change detection.In order to get the result of the high precision change detection,the change of the change pattern should be identified and eliminated.At present,the pseudo changes after change detection are recognized by the use of artificial.It is not only wasting time,spending a lot of resources,but also causing miscarriage of justice because of the identification personnel knowledge shortage.Therefore,it is necessary to build a database of eco-geographical knowledge for automatic recognition of the pseudo change after change detection.As we all known,the eco-geographical region has many characteristics,such as global,partition stability,characteristic features change regularity and large amount of information,so it can be used to construct knowledge database for change detection.Based on the eco-geographical regions,this study makes use of knowledge of ecological and geographical as well as the experience of experts to identify the pseudo change auxiliary after the remote sensing image change detection.Firstly,database of the pseudo change rule for the eco-geographical regions and the rule of land cover changes are constructed.The object-oriented method is used to design the database in order to build the inheritance relationship of rule classes,and improve the efficiency of the pseudo change recognition.Secondly,this study uses PostgreSQL to store the knowledge,and adopts C# to develop the knowledge database management system(KBMS).The recognition of the pseudo change are realized by building the pseudo-change recognition system using the eco-geographical regions knowledge database,at the same time,the accuracy of the change vector after removing the pseudo change is verified by using two test areas-Beijing and Jackson.The results show that it not only makes the automation,but also improves the accuracy of the change detection.