人工杨树林D-H曲线动态建模方法研究  

THE RESEARCH OF DYNAMIC MODELING METHOD ABOUT ARTIFICIAL POPLARS' D-H CURVE

作　　者：裴志永[1,2] 李文彬[2] 陈瑛[1] 

机构地区：[1]内蒙古农业大学能源与交通工程学院,呼和浩特010018 [2]北京林业大学工学院,北京100083

出　　处：《内蒙古农业大学学报（自然科学版）》2011年第4期168-171,共4页Journal of Inner Mongolia Agricultural University(Natural Science Edition)

基　　金：国家自然基金项目(30972425);国家林业公益性行业项目(200804029)

摘　　要：笔者研发了树木生长量远程监测系统,满足了森林精细化经营管理的需求。但是,当前所用的依据大样本量法所建立的高径曲线(D-H曲线),其与小样本动态精细监测相矛盾。本文研究运用偏最小二乘回归的方法进行D-H曲线建模,结果表明该方法能有效解决传统回归方法中依据小样本所建模型所存在的问题。建立的偏最小二乘回归方程表明自变量和因变量相应的第一主成分t1和u1之间都有较强的线性关系,相关系数达到93%和80%,且少样本组自变量第一主成分与因变量相关系数更大。结合实测值与预测值的误差检验,表明在小样本条件下依据该方法所建D-H模型有效。Refined to meet modern forest management requirements,we need to study for the amount of tree growth,remote monitoring system,height-diameter curve（DH curve）.In this paper,partial least squares regression curve modeling approach to D-H,the results show that the method can effectively solve the traditional regression method based on the model of the small sample problems.Established partial least squares regression equation shows that the corresponding independent variables and the dependent variable the first principal component t1 and u1 have a strong linear relationship between the correlation coefficient of 93% and 80%,and less variable sample group from the first a principal component and the dependent variable correlation coefficient greater.Combination of measured and predicted value of the error test,indicating that under the conditions of the small sample built according to the method effectively DH model.

关 键 词：D-H曲线 小样本 偏最小二乘回归 

分 类 号：S792.11[农业科学—林木遗传育种]