Research On Plant Modeling Method Based On Intelligent Parameter Extraction

Plant modeling has long been one of the focuses of research in the field of computer graphics.Natural plants are ubiquitous around our environment,and plants are an essential environmental factor in natural scene modeling.Moreover,plant morphological structures are complex and diverse,and many different morphological structures are required in large-scale scenarios.How to use the computer to generate models efficiently and quickly has become one of the focuses of plant modeling research.Inferring its growth model for known plants is an effective method of plant modeling.In view of the fact that the current method is not intuitive enough,the input requirements are strict and the time consumption is relatively long,this thesis studies the plant modeling method based on parameter intelligent extraction.This method can effectively achieve the formation of different forms of similar plants.The main work and results of this thesis are as follows:1.A method for calculating plant morphology similarity based on image features was proposed.The method considered the shape feature and the color feature of the plant image,wherein the shape feature included the contour feature and the regional feature of the plant image,and described the morphological characteristics of the plant from three aspects: the looseness of the plant shoots,the density of the leaves and the color.The degree of looseness of plant shoots was expressed in three aspects,the aspect ratio of the plant,the outline of the plant,and the height of the first primary branch.The degree of density of the blade was represented by the proportion of the entire enclosed rectangular area of the blade.The color of the blade was represented by a color histogram based on the HSV and YUV color space.And then it used information entropy to determine the weight.In addition,compared with the common image similarity method,the precision and recall were better,and they were not sensitive to scaling.2.A method for extracting parameter based on improved particle swarm optimization for plant growth model parameters was proposed.The method first set 8 plant growth model parameters,of which 2 parameters controled the spindle shape,4 parameters controled the side branch shape,and 2 parameters controled the growth process morphology.A description framework based on plant topology was proposed.The uniaxial and collinear frameworks were defined.The corresponding L-system rules were defined for the description framework.The parameters were optimized by using the improved particle swarm optimization algorithm combined with L-system.The experimental results showed that the fitness and standard deviation of the improved particle swarm optimization algorithm were in the leading level,and the convergence speed was also improved.Compared with the common methods,the parameter extraction method was more intuitive,required less input and took less time.3.Data processing and morphological modeling of plant growth model parameters were carried out.Data processing included data constraints and data expansion.According to the characteristics of plants,constraints were set on the data,and the Gauss distribution was used to expand the data to obtain similar data.Based on the MFC class library development system,the functions of plant similarity,parameter extraction,random morphology and simulated growth were realized.The plant modeling method based on parameter intelligent extraction proposed in this thesis provided a unified description of plant topology,which was intuitive and time-consuming.And it can provide new ideas for the study of different types of plants of the same type.It is of great significance in the filed of plant modeling.