广东二元立木材积表的编制与改进方法的研究

【作者】 刘恩斌；

【导师】 佘光辉；

【作者基本信息】 南京林业大学 ， 森林经理学， 2005， 硕士

【摘要】 森林资源的监测与估计是森林资源管理中的重要内容,森林资源的估计主要是森林(林分)的估计,单株立木材积的测定是林分蓄积估计的基础工作,单株立木材积的测定除了直接测定外,主要是按树种,测树控制因子编制的立木材积表确定立木材积,对大量林木材积的确定,材积表是一种有效的方法。 目前广泛采用的立木材积表有一元立木材积表和二元立木材积表,一元立木材积表的控制因子只有胸径,适用于区域小的范围,二元立木材积表的控制因子有胸径和树高,可适用于大区域的地区,二元立木材积表是编制立木材积表的重要基础,在很多情况下都可用二元立木材积表导算出一元立木材积表。 材积表的编制涉及多种学科如测树学,数学,计算机,生态学等。建国已来,我国对森林资源调查与林业规划投入了大量的人力,物力,材积表的编制已基本形成了较完整的体系,从方法上大部分是按采用选测一定量的树木胸径和树高,应用伐倒分段求出树木体积,再利用这三者之间的关系导算出各树种的二元材积表。 本文论述了有效建模样本的组成,分析了回归模型特点,提出了回归模型的的评价指标,并以广东森林调查资料结合生产实践中常用模型进行了分析应用。说明了非线性模型必须要用加权最小二乘算法,并论证了以模型本身构造权函数就是所要的最佳权函数。总结出编制材积表的基本步骤,并根据此步骤对广东杉木二元立木材积表进行了编制。 应用数学非线性回归理论,以marquardt非线性模型参数辩识为基础,对阻尼最小二乘算法做了改进,提出参数估计新的迭代的公式,分析了估计方法的误差,从而使材积模型的估计精度得到提高,并以山本式二元立木动态材积模型为例对改进的方法进行了检验。以经典抗差理论和最小二乘方法为基础,详细阐述了抗差因子的选取,从抵抗粗差的角度出发,提高了非线性模型在有粗差干扰条件下参数估计的精度,在此基础上提出了抗差估计的迭代公式。 对实际中常用模型进行了分析,比较,评价,从而得出性能最好的一个模型,并应用此模型编制了广东杉木二元立木材积表。 综上所述,本文对材积表编制的基本概念,基本理论,基本步骤进行了详细阐述,在此基础上又提出了编制的一些改进方法。更多还原

【Abstract】 The monitor and estimate of the forest resources are important contents in managements of the forest resources, the estimate of the forest resources is mainly the estimate of of the forest( wood cent), the estimate of volume of single stub is a foundation work in the estimated store of forest.the mensuration of volume of single stub in addition to direct measurement, is mainly to authorized tree volume according to species of tree.currently Extensive tree volumes adopted are unitary tree volume table and duality tree volume table unitary tree volume table contain a dominatedgene , duality tree volume table contain two dominated gene as tree chest and tree high. Currently, duality tree volume table may extrusionunitary tree volume table.tree volumes involves various courses such as mathematics, calculator, ecology etc.. since our nation was found, the our country plunge into a great deal of manpower, material resources to the forest resources, tree authorized volume tables come into being the system of the integrity, tree chest and tree high were measured, thus tree was fell to partition to calculate a tree physical volume, finally there relation was maked use and lead on the whole the duality tree volume tableThis paper is based on the contents above, makes the detailed introduction to the following aspect:The paper analysed organization of sample of established model, gets to set up sample of effectively established model at the scope that the wood chest may appear.Analyzed characteristics of the model of regression, puted forward to of evaluation index sign of the model of regression , investigated data with the forest of Guangdong is applied to analys and evaluate index sign of the model of regression.The paper analysed organization of sample of established model, gets to set up sample of effectively established model at the scope that the wood chest may appear.Nonlinear model must applied The least-squares Minimization Algorithm, optimal power function is maked up of model itself.The basic steps of The tree volume table summarized are applied to data with the forest of Guangdong according to this step.nonline regression theories of mathematics was applied to improve the damping least-squares Minimization Algorithm ,enter necessarily, and the estimated accuracy of t the tree volume model was improved, new iterative formula of the parameter estimated was put forward, new iterative method putted forward was carried on the examination .classic theories of nonline regression and The least-squares Minimization was applied to put forward to the least-squares Minimization Algorithm of resisting error and Resisted Dimission Error gene.according to actual things and researched result of the people of the past, the actually used model was carried on the analysis, compare, evaluate, thus get a best model.summarized, this text expound the basic concept of The tree volume table, basic theories, the basic step, on this foundation ,auther put forward some improved methods.更多还原