基于泰森多边形与人工神经网络的单木模型研究  

Research and application of GIS and ANN in individual growth model

作　　者：何瑞珍[1] 黄家荣[1] 全锋[1] 

机构地区：[1]河南农业大学林学院,河南郑州450002

出　　处：《河南农业大学学报》2009年第3期260-263,268,共5页Journal of Henan Agricultural University

基　　金：河南省科技攻关项目(0624050007)

摘　　要：利用宝天曼自然保护区栎类天然次生林6块标准地的调查数据,采用人工神经网络建模方法,以期初胸径,泰森多边形面积,期初年龄,期末年龄,缓冲区距离直径比为输入变量,期末胸径为目标变量,在MATLAB软件中进行模型训练,建立人工神经网络模型.本研究先用部分数据作为训练样本,完成模型的学习和训练,又用部分数据进行了检验,在精度满足要求时,进行了数据的模拟,由此得到了单木生长模型.这种以GIS与ANN为基础的单木生长模型能很好的预测宝天曼地区天然栎类的期末胸径,具有现实意义.The investigated data of the six standard sampling of oak natural forest was used in the Baotianman Nature Reserve, and the artificial neural network （ANN） modeling method was adopted. In the model, the beginning age, the ending age, the begining diameter at breast height, the area of the Thiessen polygon, the BDDR （buffer distance diameter ratio） were used as input variables, the end diameter was as target, the model was trained. Some other data was selected to test the model, when the precision was reached, the model was built and the end diameter was derived. The growth model based on GIS and ANN can be used to forecast the end diameter of natural oak in Baotianman areas, it is of practical significance.

关 键 词：栎类天然次生林 宝天曼 单木生长模型 缓冲区距离直径比 人工神经网络模型 

分 类 号：S792.18[农业科学—林木遗传育种]