Gpu-based Virtual Plant Growth Study

"Virtual plant growth"is computer simulation of the structural development and growth of plants in 3D space. It can be widely used in natural scenery simulation, ecology, agriculture and so on. The simulation of plant-growth is an important content in the research of"Visual plant growth". Currently, there are several factors limiting the plant photorealistic graphics rendering speed, such as morphological diversity, complex plant details and numerous computer graphic elements required in computer simulation of plants in 3D space. How to accelerate the plant morphology rendering speed becomes a hot topic for researchers.In order to accelerate the plant morphology rendering, we study on the following two aspects. One is making full use of high-performance hardware, such as Graphics Processing Unit (GPU). The other is constructing a new plant morphological structure model to achieve fast rendering. The main contents and innovations are as follows:(1) Dual-scale automaton model based on plant-growth principles is a vivid and structural model. This model can be used to effectively simulate the plant growth. In order to accelerate the rendering speed of the model, a GPU-based dual-scale automaton method is proposed. The 3D structure of the plant is constructed by topological structure and normalized plant organs. Topological structure is firstly constructed in vertex shader, and then geometrical structure is built in geometry shader according to the topological structure, geometric parameters and normalized plant organs. Finally, geometrical structure is transferred to frame buffer for displaying. In this paper, we give some examples. The results demonstrate the effectiveness of the rendering method and the remarkable improvement of the speed.(2) In order to simulate the morphology structure of plant, a plant morphological structure model is proposed. The plant morphological structure is described from two aspects: topological structure and geometrical structure. Geometrical structure is constructed by topological structure and normalized plant organs. Metamer is used as the unit for describing the morphological structure of plant, which helps keep the structure details. Growth characteristic is chosen as the standard to classify the characters of plant morphological structure, so that the simulation data required for modeling plant morphological structure can be compressed. This model is tested in this paper, and the results demonstrate the effectiveness and feasibility of the model.The GPU-based rendering for dual-scale automaton model and the plant morphological structure model can be applied to natural scenery rendering, plant visualization, computer-aided design etc, which involve enormous data.