文献类型：期刊文献

中文题名：基于人工神经网络对栎树天然林地位指数模拟系统的研究

英文题名：Study of oak growth dynamic simulation system based on artificial neural network

作者：黄旭光[1,2] 周俊朝[1] 黄柏华[2] 黄家荣[1] 杨保国[2] 谢鹏芳[1]

第一作者：黄旭光

机构：[1]河南农业大学林学院;[2]中国林业科学研究院热带林业实验中心

年份：2015

卷号：49

期号：2

起止页码：190-194

中文期刊名：河南农业大学学报

外文期刊名：Journal of Henan Agricultural University

收录：CSTPCD;;北大核心:【北大核心2014】；CSCD:【CSCD_E2015_2016】；

基金：河南省科技攻关项目(0624050007)

语种：中文

中文关键词：薄山林场;栎树;人工神经网络;地位指数模拟系统

外文关键词：Boshan forest farm; oak tree; artificial neural network; site index simulation system

分类号：S757

摘要：以驻马店薄山林场的天然栎树次生林为研究对象,用人工神经网络建立地位指数模型。用MATLAB语言编程,开发出栎树天然林地位指数模拟系统。以10块栎树标准地的调查数据为训练样本,对所建模型进行分析、训练。结果表明,最佳网络结构为1∶3∶1,其拟合精度为98.36%;系统模拟完全符合森林资源调查及预测的精度要求,是一个结构简单、界面直观、操作方便的地位指数模拟系统。
The site index model was established based on artificial neural network,taking the natural secondary forest oak tree in Boshan forest farm of Zhumadian as research object. By using MATLAB programming language and based on the established site index model,an automatic query interface was developed,and a natural oak forest site index simulation system was developed. Using the survey data from 10 oak standard plots as training samples,the established model was analyzed and trained. The results show that the optimal network structure is 1∶ 3∶ 1,and the overall fitting accuracy is 98. 36%.The simulation system is fully in compliance with the accuracy requirements of forest resource survey and prediction,with simple structure and intuitive interface,and is easy to operate.