Automatic 3D-Visualization of Geo-Information for Fireprevention in European Mediterranean Areas

The providing of basic data for decision procedures in disaster management is carried out to an ever increasing extent by means of remote sensing methods. Here, not only the accuracy and availability of data, but also an easy and affordable evaluation represents an important factor. That is why the automation of all sorts of processes in data preparation and analysis gains increasingly more importance. Within the framework of this contribution it will be described how basic data are acquired and visualized automatedly by means of remote sensing methods for areas, where existing maps and geodata are available only to a small extent. Here, a continuous production line starting from the creation of an elevation model, geo-referencing and classification up to the automated 3D-visualization of test areas in the Mediterranean area is described. The data processing ranges from the calculation of surface models, digital elevation models (DEMs) and geocoding of high resolution satellite images to classification of vegetation and biomass calculation. Based on these DEMs as well as on optimized parametric sensor models, multispectral as well as panchromatic Quickbird images are geocoded. The classification of vegetation makes use of “pan-sharpened” Quickbird images resulting from data fusion of panchromatic and multispectral images. Biomass data can be calculated from the classification results and forest heights, which are extracted from surface models in order to support fire prevention carried out by local forest managements. For the presentation of these results 3D landscape visualization plays a crucial role because classification results which may be raster or vector maps using specific colours for different classes are not easy to understand for all users involved into fire prevention. The construction of photo-realistic 3D views and 3D animations based on artificial textures and on artificial ecosystems allows us to realise an easily understandable presentation of the current vegetation situation and possible changes in the future. Thus the automatic 3D visualization of raster and vector data will be the main aspect of this paper. Based on the landscape model constructed from the above described DEMs and classification results, a strategy will be described making possible the automatic 3D visualization of every vegetation analysis and change simulation without additional effort. Thus the fire fighting services will get an additional innovative tool for planning and decision making.