3D Visualization of the Watzmann-Massif in Bavaria of Germany

Transcription

1 3D Visualization of the Watzmann-Massif in Bavaria of Germany Angsüsser, St. 1 & Kumke, H. 2 Department of Cartography, Technical University of Munich, Arcisstraße 21, D München, Germany, angs@bv.tum.de 2 holger.kumke@gmx.de Abstract The paper deals with 3D cartography. Using the program 3D studio max (Autodesk) a flight over an alpine region ("Watzmann-Massif") is animated. In addition to a description of the project some thoughts and ideas especially about generalisation degree are discussed. Keywords: 3D cartography, generalisation, animation, flight, Watzmann-Massif (Bavaria, Germany) Introduction The presented project deals with possibilities of 3D cartography. For that purpose a flight over Watzmann-Massif (Bavaria, Germany) is animated (Kumke 2001). By the term 3D cartography we mean all cartographic terrain presentations that explicitly include the third dimension (z-axis). Such visualisations are based on 3D models. In analogous cartography there was no difference between these models and the map itself (Petrovic, 2001). In modern cartography it is possible to create several different 3D presentations from one single 3D model. Depending on the method how this is done, we make a difference between real 3D, pseudo 3D, dynamic 2D and static 2D (Buchroithner & Schenkel 2001). The film produced in this project is an example for a dynamic 2D visualisation. There are also static 2D presentations in form of perspective images. The availability of more depth cues in such films leads to a higher immersion degree, hence a more realistic 3D impression than in images. Goals of this project This project has a two-fold goal. On the one hand, the film and the images are created to explore the potentialities of 3D studio max (Autodesk) for 3D cartography. On the other hand, the combination of different abstraction levels will be tested. That is to say, our final product should be adaptive to different user groups and their varying needs. Project description Basic data As basic data served the digital elevation model (DEM) and the topographic map 1 : (TK25, Fig.1). Both were available for the chosen region (size: 11x12km) from the Surveying and Mapping Agency of Bavaria (Bayerisches Landesvermessungsamt). The resolution of the DEM is 50m which corresponds to height points.

2 The digital TK25 contains five layers that are digitized colour separations of the analog map with a resolution of 508dpi (Tab.1). The brown elements like contour lines and contour values were not used because they do not fit the DEM due to their higher precision than the latter. However there are some parts of differently coloured contour lines remaining in other layers (e.g. contours in glaciated areas, rock contours). Unlike the horizontal run of a "normal" contour, they typically undulate along the slopes. A special kind of contours, depth contours, formed an exception. There was no problem to integrate them in the map overlay because the underwater morphology of the Königssee (King Lake) is not shown. Water surface is visualised as an opaque area instead. Within the chosen region fifteen buildings were selected to be modelled in more detail. For this purpose following data was additionally surveyed: By using a digital camera (Kodak DCS 420, resolution: 1012x1524 pixel) several photos were taken to get an idea of size, shape and look of the different houses as well as the church St.Bartholomä. Some texture-like details have also been recorded. Furthermore a small number of measurements were made to have a clue for size fixing. Surveying and Mapping Agency of Bavaria Fig.1: A section of the TK25 (approximately original size) colour black cyan cyan (half tone) brown green Elements planimetric presentations, rock drawing, rock contours, text streams, watersides, depth contours, contours in glaciated areas, text water surfaces contour lines, contour values Wood Tab.1: The five colour separations of the TK25 (the brown layer was not used) The described data had to be imported into 3D studio max. For the DEM this was done using a plug-in named "terrain 2" (free download: The four layers of the TK25 were coloured before they could be merged to one map. After resolution reduction from

3 508 to 300dpi the overlay was imported into the program (as.tif). In addition, the digital photos were integrated (different formats like.tif,.bmp or.jpg are possible). Modelling Before modelling can be started, a desired abstraction degree (Level of Detail = LoD) has to be determined. Possible presentations range from pure geometric to photorealistic ones (Tab.2). Which presentation should be selected depends mainly on the purpose of the visualisation, the expected type of users and the quality of available data. As stated above one of our goals was to explore the potentialities of 3D studio max. For this reason a high LoD was desirable in spite of an abstract map overlay. Criterion pure geometric presentation photorealistic presentation Origin artificial artificial natural LoD low high infinite generalisation degree e.g. abstraction degree typification degree high low none 1 nature (original scene) individualisation degree low high only individuals time dependence low high complete information perception little (selected) much information infinity of information information in in long time in infinity of time 2 short time 1 there is a kind of natural generalisation based on our cognition 2 the process of perception is terminated if the searched information was found or the process is judged as inefficient or not promising Tab.2: Comparison between different kinds of cartographic presentations and the original Besides fifteen chosen buildings, two ships, two landing stages and some trees were modelled (Fig.2). The modelling process requires that the object geometry be defined. The size and shape of the selected objects were determined with help of the surveyed measurements and photos. For the church a simple ground plan was additionally used. The object surface was constructed by integrating the textures (so called "maps"). Many standard textures are included in the program. An option is available to create individual textures as well. In this context the photos have proved very helpful.

4 (a) (b) Fig.2: The church St.Bartholomä: (a) original scene and (b) model Animation Every animation consists of three object types (Dransch 1997): graphical objects, camera and light source. At least one of them has to be dynamic. Other objects can be either static or dynamic (Tab.3). Camera is always one of the dynamic objects in flight simulations. In addition two ships and a text sample were animated. Text concerns the name of the massiv "Watzmann". It changes in size and rotates depending on the position of the camera. static objects DEM with map overlay Buildings Trees landing stages light sources Sky fog dynamic objects camera ships text Tab.3: Static and dynamic objects in the animated flight Flight simulations are typical examples for temporal animations. There exists a direct relation between display time and world time. The described film was computed in advance. For this reason it can be assigned to the "real-time later" animation type (Kraak & Ormeling 1996). The most important advantage of this type is the possibility to work with huge amounts of data as they usually occur in photorealistic presentations. Although the film is just, to some extent, near to photorealism and lasts only 1:40 minutes it needs 160MB memory as audio visual interleaved (.avi) file. A great disadvantage of this animation type is the absence of interactivity (Dransch 1997). Hence, it is impossible for the user to choose an individual flight route. Discussion of examples In addition to the dynamic camera other static cameras were used to take pictures that represent the landscape at selected moments. These perspective views (static 2D presentations) of the model can give an idea about its nature (Fig.3 and Fig.4).

5 (a) Fotokunstverlag F.G. Zeitz KG (b) Fig.3: The Watzmann-Massif (2713m), the central part of the Königssee (603m) and in the midst of them the alluvial cone where the modelled buildings are situated: (a) original scene and (b) model (a) Fotokunstverlag F.G. Zeitz KG (b) Fig.4: The church St.Bartholomä at the waterside of the Königssee and the surrounding buildings: (a) original scene and (b) model It is obvious that the model is inhomogeneous. - whilst sky and clouds are photorealistic and buildings are nearly photorealistic, the map overlay is much more abstract. - whilst terrain and buildings are 3D, vegetation (except some solitary trees) and clouds are 2D - whilst buildings are modelled in original size, the symbols on the map are bigger as a result of generalisation. Like some photo- and satellite-maps inhomogeneous presentations can lead to confusion. Because of the higher immersion degree this effect is also more distinctive than in 2D maps. To get a more homogeneous kind of visualisation, it is necessary to present all involved object

6 types (terrain, buildings, vegetation, sky and surface) in a uniform way. This means that dimensionality, geometry (above all size and shape) and texture should be equally treated regarding their generalisation degree / LoD. As a matter of fact, there is no completely free choice to visualise a landscape in a homogeneous way. It hardly makes sense to present terrain as abstract 2D hypsometric model in 3D cartography. So it is predetermined to use a 3D model for that purpose. As a consequence all object types that explicitly include the third dimension (mainly buildings, vegetation) should be presented like terrain with an explicit z-axis. In fact the problem is even more complicated because of the continuously changing scale. "Because an unobtrusive symbology is difficult to achieve in perspective, 3D cartographic visualisations tend towards naturalism" (Swanson 1999:183). This has to be confirmed, but it cannot be the only goal in 3D cartography to reach photorealism. As Haeberling stated in the same year, these problems are part of a "new branch of mapping theory" (Haeberling, 1999). Before this branch can grow adequately, maybe photorealism is needed first to realise the advantages of a higher abstraction degree. Especially for interactive real-time solutions with integrated hyperlinks a lower LoD makes navigation easier and more comfortable. Conclusions and Perspectives It was shown that the program 3D studio max offers many possibilities for 3D cartography. There is no problem to integrate different overlays, use or create various textures, model and animate objects or add cameras and light sources. In short: the tools are ready but we have to know how to use them most efficiently. Yet there is a lack of basic research in 3D cartography. The terms "map users" and "map perception" (Haeberling, 1999) still remain as hot topics. A more user-centered cartography is suggested by many cartographers. The acquired knowledge about user s perception and behaviour is far from being sufficient. The implementation of adaptive visualisation strategies in 3D cartography highly depends on advances in these fields. No doubt that 3D cartography offers interesting possibilities for visualising and exploring a terrain. The close relationship between nature and map impression leads to a wider range of potential users. People who normally have problems with 2D terrain presentation are now able to navigate in the world of 3D cartography. This development can also be seen as a further step towards a more user-oriented cartography. Development of an adequate (referring to purpose and user), homogeneous generalisation degree is a special challenge in 3D cartography. Some other problems are associated with this question. It is obvious that the impression of a homogeneous generalisation degree can only be achieved if the generalisation is interrelated to the inhomogeneous scale. For this reason different approaches are tested to reach a kind of continuously zigzagging generalisation degree (e.g. Chen G. et al. 2001, Döllner 2001). In this context efforts have been made to find an appropriate system of 3D symbols (e.g. Haeberling 1999, Bandrova 2001, Chen X. & Bai 2001).

7 References Bandrova, T. (2001): Designing of Symbol System for 3D City Maps. In: Proceedings of the 20th International Cartographic Conference, Beijing, 2001, Volume 2, Buchroithner, M.F. & Schenkel, R. (2001): Kartographische Echt-3-D-Visualisierung - eine Ergänzung zu flachen Karten. In: Buzin, R. & Wintges, Th. (Hrsg.): Kartographie multidisziplinär und multimedial. Beiträge zum 50. Deutschen Kartographentag. Herbert Wichmann Verlag, Heidelberg, 2001, Chen, G.; Wan, G. & Wu, Z. (2001): An Integrated Global Visual Simulation System Based on Map Data. In: Proceedings of the 20th International Cartographic Conference, Beijing, 2001, Volume 4, Chen, X. & Bai, F. (2001): Generalization of Three-Dimensional City Maps. In: Proceedings of the 20th International Cartographic Conference, Beijing, 2001, Volume 3, Döllner, J. (2001): Informationsvisualisierung mit dynamischen, interaktiven 3D-Karten. In: Kartographische Nachrichten, 51.Jg., Heft 4, Kirschbaum Verlag, Bonn, 2001, Dransch, D. (1997): Computer-Animation in der Kartographie. Theorie und Praxis. Springer, Berlin, 1997, Haeberling, Ch. (1999): Symbolization in topographic 3D-maps: conceptual aspects for useroriented design. In: Proceedings of the 19th International Cartographic Conference, Ottawa, 1999 (published on CD-ROM). Kraak, M.J. & Ormeling, F.J. (1996): Cartography: Visualization of Spatial Data. Addison Wesley Longman, Harlow, 1998 (reprint), Kumke, H. (2001): Demo "Flug über das Watzmann Massiv". In: Hake, G.; Grünreich, D. & Meng, L.: Kartographie Visualisierung raum-zeitlicher Information. Walter de Gruyter, Berlin (attached CD-ROM) (to be published). Petrovic, D. (2001): Three-Dimensional Mountain Map. In: Proceedings of the 20th International Cartographic Conference, Beijing, Volume 5, Swanson, J. (1999): The Cartographic Possibilities of VRML. In: Cartwright, W.; Peterson, M.P. & Gartner, G. (eds).: Multimedia Cartography. Springer, Berlin - New York, 1999,