Visual Modeling and Simulation of Plant Growth Based on Single Image

Abstract

Fractal is an important theoretical basis in the visualization research of virtual plant growth modeling. Fractal theory is an active branch of mathematics, and its research object is non-smooth and non-differentiable highly irregular geometric objects produced in nonlinear systems. In this paper, a plant leaf simulation method based on single image and botany theory is proposed. Combining DLA (Diffusion Limited Aggregation) model with botany's theory that the growth of vein is influenced by auxin and leaf shape, vein is automatically generated as leaf skeleton. Scan plant leaves with scanner to obtain leaf point cloud data. The Laplacian operator of Gauss weight graph of each point in the point cloud is calculated, and the alternative seed point set is extracted according to the module length of the operator. A plant growth model based on state machine is proposed, and random vectors are introduced to predict the growth of plants. The simulation results show that the three-dimensional model constructed by this method has good reducibility to the true shape of grape leaves, and the growth model can simulate and realize the visualization of grape branches. This study is of positive significance to the 3D visualization of crops based on geometric models.