当前位置: 首页 > 文章 > 基于5年生火炬松建立木材基本密度近红外预测模型 福建农林大学学报(自然科学版) 2021 (6) 767-770
Position: Home > Articles > 基于5年生火炬松建立木材基本密度近红外预测模型 Journal of Fujian Agriculture and Forestry University(Natural Science Edition) 2021 (6) 767-770

基于5年生火炬松建立木材基本密度近红外预测模型

作  者:
冯志恒;吕欣欣;李赛楠;周晓煦;蒋开彬;黄少伟
单  位:
华南农业大学林学与风景园林学院/广东省森林植物种质资源创新与利用重点实验室
关键词:
近红外模型;基本密度;火炬松;
摘  要:
以5年生火炬松为研究对象,采用DA7200近红外光谱成分分析仪扫描火炬松木芯,利用近红外光谱数据,结合偏最小二乘法建立5年生火炬松木材基本密度近红外快速预测模型.比较分析不同预处理模型评价的相关参数,得到一阶导数与平滑算法结合的最佳预处理方法,校正集相关系数达到0.925 2,校正集均方根误差为0.005 7 g·cm-3,交互验证集相关系数达到0.796 2,交互验证集均方根误差为0.009 5 g·cm-3,外部验证集预测值与实测值具有较高的相关性(R=0.805 7),预测均方根误差为0.013 9 g·cm-3.本研究建立的模型能满足对5年生火炬松木材基本密度的快速预测.
作  者:
FENG Zhiheng;Lü Xinxin;LI Sainan;ZHOU Xiaoxu;JIANG Kaibin;HUANG Shaowei;College of Forestry and Landscape Architecture, South China Agricultural University/Guangdong Key Laboratory for Innovative Development and Utilization of Forest Plant Germplasm;
关键词:
near-infrared model;;basic density;;loblolly pine
摘  要:
To develop a rapid density prediction model based on near infrared spectroscopy, wood cores of 5-year-old loblolly pine were scanned by DA7200 near-infrared spectroscopy component analyzer, then a wood density calculation model was proposed combined with partial least squares method. After comparing parameters generated from different preprocessing models, an optimal preprocessing model combining functions of first order derivative and smoothing algorithm was selected, with the correlation coefficient of the calibration set reaching 0.925 2 and root mean square error(RMSE) reaching 0.005 7 g·cm~(-3). The correlation coefficient of the cross validation set arrived at 0.796 2, with the RMSE being 0.0095 g·cm~(-3). Furthermore, high correlations(R = 0.805 7) and low RMSEP of 0.013 9 g·cm~(-3) were found between the predicted values and measured values by external validation, indicating the accuracy and reliability of the preprocessing model.

相似文章

计量
文章访问数: 9
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊