Abstract:
Traditionally, GIS have used data generated from remote sensing, but only after the data has been pre-processed in some way to provide a suitable classification and/or segmentation. This leads to several weaknesses, the resulting system is inflexible and cannot support multiple (possibly overlapping) segmentations of the same area. We describe a new GIS under development that includes a set of image understanding algorithms, that work to extract features of interest from the raw images. The algorithms employed are chosen by the system to best emphasise the types of features (rivers, fields, forests etc.) that the user is currently investigating. The GIS model has been extended to allow multiple representations of the same features to co-exist, enabling the image segmentation to be adaptive and not fixed.