Abstract :
Most landscapes are covered by some type of vegetation, therefore digital modelling of vegetation is a prerequisite for three-dimensional landscape visualisation. Vegetation can be modelled either at the level of individual plants or as a terrain texture.
At object scale, plants can be described by a set of rules which define the shape of its individual parts (stems, leafs, flowers, etc.) as well as their configuration and temporal development. Some applications available for planners are based on botanical models, such as L-grammar [Development models of multicellular organisms: a computer graphics perspective. In: Langton, E.C. (Ed.), Artificial Life. Addison-Wesley, Reading, MA, 1988, p. 221] or the AMAP system [J. Biosci. 17 (3) (1992) 275]. The use of this approach often yields very large files and requires powerful computing equipment. Other applications aim at an optimisation of the rendering speed by using simplified, more symbolic objects.
At landscape scale, vegetation is often described as a texture of the terrain. This can be done by simple image mapping or by three-dimensional texture mapping techniques, such as bump mapping, displacement mapping or particle emission mapping.
The most difficult scale level is the one between object and landscape scale, where both approaches have to be combined in order to optimise the visual quality as well as the processing speed.
In the market, an increasing number of plant visualisation systems is now available. Selection criteria from the point of view of planning practice should not only include aspects of rendering quality, but also aspects of integration with standard applications at object and landscape scale, such as CAD, GIS, DBMS, and terrain modelling.
