基于疤痕大小建立节子体积预测模型  

Establishment of a volume prediction model for nodes based on scar size

作　　者：李泽霖 贾炜玮[1,2] 纪鑫妍 郭昊天 赵阳[1,2] LI Zelin;JIA Weiwei;JI Xinyan;GUO Haotian;ZHAO Yang(College of Forestry,Ministry of Education,Northeast Forestry University,Harbin 150040,Heilongjiang,China;Key Laboratory of Sustainable Management of Forest Ecosystem,Ministry of Education,Northeast Forestry University,Harbin 150040,Heilongjiang,China)

机构地区：[1]东北林业大学林学院,黑龙江哈尔滨150040 [2]东北林业大学森林生态系统可持续经营教育部重点实验室,黑龙江哈尔滨150040

出　　处：《中南林业科技大学学报》2024年第4期66-74,共9页Journal of Central South University of Forestry & Technology

基　　金：国家自然科学基金区域创新发展联合基金重点项目(U21A20244);国家自然科学基金面上项目(31870622);中央高校基本科研业务费专项资金项目(2572019CP08)。

摘　　要：【目的】以黑龙江省林口林业局、东京城林业局林场29块标准地中49株人工红松共计1207个节子数据为研究对象,并以其相关的数据为基础,构建节子体积预测模型,通过人工红松树干内部节子体积的分布规律分析,实现无破坏性地评价木材质量,为科学合理地利用木材提供参考。【方法】首先通过剖析法收集相关数据,查阅相关文献,建立基础模型;在构建混合模型时,引入样木与样地水平的随机效应,通过比较赤池信息准则AIC、贝叶斯信息准则BIC等评价指标,选出拟合效果最佳的混合模型;构建双水平嵌套混合模型时,将样木水平嵌套于样地水平中,根据拟合评价指标,得出最优的双水平嵌套混合模型。【结果】3种模型中,单水平混合模型与双水平嵌套混合模型的拟合精度均大于基础模型,AIC、BIC等评价指标显示,双水平嵌套混合模型在对节子属性的拟合效果更优于单水平混合效应模型。检验结果显示,基础模型的预测精度大于90%,两种混合模型的预测精度均在98%以上,说明构建的模型能够对节子体积进行较好地预测。【结论】基于双水平嵌套混合模型,不仅能实现无破坏性地预估节子体积,还能更加准确地体现人工红松树干内部节子体积的分布规律。实现无破坏性地评价木材质量,为科学合理地利用木材提供参考,进而提高经济收益。【Objective】This study took 49 artificial Pinus koraiensis with a total of 1207 knot data from 29 standard plots in the Linkou Forestry Bureau and the Dongjingcheng Forestry Bureau in Heilongjiang province as the research object,and a knot volume prediction model was constructed based on their relevant data.Through analyzing the distribution law of the internal knot volume of the artificial Korean pine trunks,it achieved non-destructive evaluation of wood quality and provided a reference for the scientific and rational utilization of wood.【Method】Firstly,we collected relevant data through analysis and reviewed relevant literature to establish a basic model.When constructing a mixed model,introduced random effects at the sample tree and sample plot levels,and selected the best fitting mixed model by comparing evaluation metrics such as AIC and BIC;When constructing a two-level nested mixed model,sample trees were nested at the sample plot level,and the optimal two-level nested mixed model was obtained based on the fitting evaluation index.【Result】Among the three models,the fitting accuracy of the single-level mixed model and the two-level nested mixed model was higher than that of the basic model.The AIC,BIC,and other evaluation indicators showed that the two-level nested mixed model had a better fitting effect on the attributes of knots than the single-level mixed effect model.The test results showed that the prediction accuracy of the basic model was greater than 90%,and the prediction accuracy of both hybrid models was above 98%,indicating that the constructed model could provide good predictions for nodule volume.【Conclusion】Based on the two-level nested mixed model,not only can non-destructive estimation of knot volume be achieved,but also the distribution law of knot volume inside the artificial P.koraiensis trunk can be more accurately reflected.Realize non-destructive evaluation of wood quality,provide reference for scientific and rational utilization of wood,and thus improve economic benefi

关 键 词：红松人工林 节子体积 疤痕面积 混合模型 

分 类 号：S781[农业科学—木材科学与技术]