Case Study: Habitat Mapping
By using the thematic relationships between the level 1 and level 2 classifications and the spatial relationships of the objects LandBase may be recoded towards specific user requirements such as high resolution habitat maps.
The example in figure 3 shows how by building logical queries specific classes can be achieved with only basic building blocks. For example regions of salt marsh were defined as herbaceous vegetation intersecting aquatic areas and also within a coastal zone. Separating broad grass types was also derived in a similar way whereas areas of scrub can be defined by looking for brad regions of scattered shrubs. Built up area (BUA) was also masked out of the map by not only taking buildings and artificial surface but also growing out to include gardens. Such automatic recoding may form the basis for further work using expert image interpretation or combining with other datasets such as satellite imagery.