基于L-M算法的林分径级材种出材率BP网络模型  

BP Neural Network Model for the Diameter Class Timber-produced Rate Based on L-M Optimization Algorithm

作　　者：郑德祥[1] 黄传春[2] 谢益林 涂勇鑫 黄斌 

机构地区：[1]福建农林大学林学院,福建福州350002 [2]福建省林业厅资源总站,福建福州350005 [3]永安林业集团股份有限公司,福建永安366000

出　　处：《北华大学学报（自然科学版）》2008年第5期461-465,共5页Journal of Beihua University（Natural Science）

基　　金：福建省青年创新人才项目(2006F3015)

摘　　要：以杉木人工林伐区设计书与检尺码单资料为基础，选择平均胸径、平均树高、立木密度、每hm^2蓄积量、郁闭度、林龄、立地类型为输入神经元，应用L-M算法对各径级材种出材率进行BP网络模型模拟预测及精度检验．结果表明，基于L-M算法的林分材种出材率BP神经网络模型具有较高精度，可为林分经验材种出材率表的编制提供一种新的思路与方法．A network topological structure was determined and an artificial network model for predicting the timber-produced rate of each diameter class was established by using L-M optimization algorithm and the precision of BP neural network model was tested according to the felling area design of Chinese fir plantation and the data of timber-scaling code list,and in which the input nerve cells include average dbh,average height,stem density,the growing stock per hectare,crown density,forest ages and site types. The result shows that the BP neural network model for the diameter class timber-produced rate based on L-M optimization algorithm has higher precision so that it can be served as a new method for the compilation of the timber-produced rate table.

关 键 词：杉木 伐区设计 检尺码单 BP网络 材种出材率 

分 类 号：S711[农业科学—林学]