A study of virtual plant is one of the emerging research fields in computer graphics, information agronomy and applied mathematics. Fast rendering and botanically faithful description of plants are a real challenge in the study. From botanical point of view, photosynthesis and organogenesis are main functions for plant growth. Photosynthesis made by leaves generates biomass that is shared by different organs according to total demand and relationships between sources and sinks of trees. Different organs take part of biomass, and thus plant grows. All of such intrinsic performances are shown through plant topological structure which is the representation of physical link among different organs and internal functions for the whole plant structure. How to simulate both functions and compute plant production based on certain plant topological structure is very important for realistic tree growth modeling. Usually, plant production is computed using the method internode by internode, while there are a lot of internodes in an individual tree, therefore, this approach is quite time-consuming even for a medium-size tree. Thus, we developed a new plant architecture and function model based on substructure decomposition. The main contributions in this dissertation are the followings: i. A new method based on substructure decomposition was put forward for fast plant description and thus for fast plant topological structure construction corresponding to different plant architectural models. This algorithm concerns all the plant architectural models, which bear instances with self-similarity, so this algorithm could integrate the characters of both L-system and fractal. Using this algorithm, we can save time on computing plant production and thus constructing plant architectural models. The higher the complexity of the plant (i.e. branching density and depth) is, the more efficient the new method is. Therefore, this method presents significant benefits in visualization of both complex structured plant and can be applied to describe plant population. In addition, this method provides a new idea for LoD technique in the application of fast plant visualization method from computergraphics point of view. 2. A new general plant architecture and function model, namely, GreenLab, in which both plant topological model and hydraulic model are well integrated, was developed based on the botanical knowledge and substructure decomposition to simulate plant growing process. Regarding growth cycle as time varying scale, physiological age as different metamorphosis stages, fresh matter and geometrical size of plant organs as space varying scale, substructure of different physiological ages at different chronological ages as different level of detail for visualization, GreenLab model can not only be faithful to plant intrinsic developing rules but also fast and simply show people the deta
修改评论