Intelligent data fusion for environmental monitoring
David G. Goodenough, P. Bhogal, Daniel A. Charlebois, Stan Matwin, Olaf Niemann · 2002
In order to monitor the land environment, it is necessary to compare past and present states and to fuse remote sensing data from multiple sources. The past states of the vegetation for test sites in British Columbia are represented in GIS files and in past remote sensing imagery. The present state is shown in current airborne and satellite imagery. The data fusion involves radiometric calibration, geocoding, and topographic relief correction of the remote sensing data, integration of these data with the multiple GIS files, and analysis with the fused data. These operations and the subsequent analysis require complex software components. These software components require advanced expertise to correctly perform the data fusion operations. A methodology for creating a sequence of intelligent expert systems has been devised and applied to the problem of data fusion. An example will be given of the application of this intelligent system.