基于样本库的杉木林分生长动态可视化模拟

Dynamic Visual Simulation of Chinese Fir Stand Growth Based on Sample Library

[目的]通过构建杉木生长样本库,研究一种考虑林分空间结构与生长之间交互过程的生长模拟方法,同时结合三维可视化技术,实现林分生长动态可视化模拟。[方法]以2012—2017年湖南省黄丰桥国有林场6块杉木样地的连续调查数据为数据源,考虑相邻木树冠侧方挤压及上方光照遮盖对中心木生长的影响,研建基于空间结构单元的水平、垂直空间结构参数,在此基础上,选取特征指标构建样本库,以马氏距离作为待模拟林木与样本库的匹配准则,对林木与样本库进行迭代匹配,实现考虑空间结构与生长交互过程的林分生长模拟。同时基于三维模型剖分及动态组织方法,采用节点动画技术对林分生长动态进行可视化表达。[结果]经相关性对比分析,选取年龄、胸径、水平、垂直空间结构参数及胸径连年生长率作为特征指标构建样本库预测胸径生长(包含杉木8 934株,其中死亡97株),采用线性建模预测树高、冠幅、活枝下高生长。通过对样地观测值与生长预测值进行配对样本T检验及线性回归,结果表明:胸径、树高、冠幅、活枝下高的生长预测效果良好,具体顺序为胸径>树高>冠幅>活枝下高,线性回归决定系数R^2分别为0.925、0.556、0.482、0.459。[结论]基于样本库的林分生长动态可视化模拟方法能实现胸径、树高、冠幅、活枝下高4个指标的生长可视化,效果形象直观,能满足林业科研和生产实践要求。

[Objective] By constructing the Chinese fir growth sample library to study a method of growth simulation considering the interaction process between spatial structure and growth of the stand. At the same time, combined with three-dimensional visualization technology, try to realize the dynamic visualization of stand growth.[Method] Taking the continuous survey data of 6 plots of Huangfengqiao State-owned Forest Farm in Hu’nan Province from 2012 to 2017 as data source, in consideration of the influence of lateral compression of the adjacent crowns and the effect of the upper light cover on the growth of the central tree, the horizontal and vertical spatial structure parameter based on spatial structure unit were developed, then the characteristic indexes were selected to construct the sample library, and using Mahalanobis distance as the matching criterion of sample library and wood to realize the simulation of stand growth considering the interaction between spatial structure and growth. Node animation was used to realize dynamic visual simulation of stand growth based on the 3D model segmentation and dynamic organization.[Result] Through the comparative analysis of correlation, the authors selected the age, DBH (D), horizontal (PH), vertical spatial structure parameters (PV) and Annual growth rate of DBH as the characteristic indexes to build the sample database to predict the growth of DBH (including 8 934 Chinese fir and 97 dead), and linear modeling was used to predict the growth of tree height (H), crown width (CW) and height under living branch (UBH). The result of paired T-test and linear regression between observation and prediction showed that the sample library had a good effect on the growth prediction of D, H, CW and UBH. The specific order was D>H>CW>UBH, determination coefficient of linear regression (R^2) were 0.925, 0.556, 0.482, and 0.459.[Conclusion] The method of dynamic visualization of stand growth based on the sample library can realize the growth visualization requirement of four indicators of D, H, CW and UBH, and it’s effect is intuitive, can meet the requirements of forestry scientific research and practice.