Image classification with LiDAR and GIS-data: moving from land cover to land use

摘要

Using image and LiDAR data the identification of image objects can be improved over using image data alone but the classification will still be mainly on the level of land cover. In order to move from land cover to land use it is suggested to include data stored in databases which can supply some information on land use. Although some of these will be stored in a GIS and can thus be linked more easily, others will typically only be indexed by street addresses with very little other geographical references. With the availability of georeferenced postal addresses these data can now be linked to image objects and used for classification purposes. An example is presented where a land cover classification is linked to the information taken from the yellow pages. This allows the identification of buildings used for commercial as well as public services down the specific branches of business. Inversely this also allows the identification of all residential buildings or those that have some kind of mixed use.