摘要
准确估算叶绿素含量对于植物生长监测、产量预测、生境的适宜性评价具有重要作用。为寻求叶片叶绿素含量的高精度估算模型,以石楠为对象,实测叶片叶绿素含量和反射光谱反射率,对原始光谱进行变换并计算植被指数,通过相关性分析挑选特征波段,运用多元逐步线性回归和偏最小二乘回归建立叶绿素预测模型。结果表明:1)FDR的逐步线性回归模型和偏最小二乘模型优于R、1/R、LR、SDR;2)DNDVI(R_(645),R_(1370))的指数函数模型为估算叶绿素含量的最佳单变量模型;3)DRI(R_(747),R_(1464))与RI(R_(733),R_(944))的逐步线性回归模型精度最高,验证结果的决定系数R^(2)为0.955,均方根误差RMSE为3.145。因此,该模型可以实现叶片叶绿素含量的准确估算,从而为实现高光谱技术监测植被叶绿素含量变化提供依据。
Accurate estimation of chlorophyll contents plays important roles in plant growth monitoring,yield prediction and habitat suitability evaluation.In order to find a high-precision estimation model,Photinia phoebe was taken as the research object to measure the chlorophyll content and reflectance of the leaves.Firstly,the original spectra were transformed and the vegetation indexes were calculated.Secondly,the characteristic bands were selected through correlation analyses.Then the chlorophyll prediction models was established by stepwise linear regression and partial least square regression(PLSR).The results showed that 1)the stepwise linear regression model and partial least square model of FDR were better than R,1/R,LR and SDR.2)The exponential model of DNDVI(R_(645),R_(1370))was the best single variable model for estimating chlorophyll contents.3)The stepwise linear regression model of DRI(R_(747),R_(1464))and RI(R_(733),R_(944))was the most accurate model,and the validation results showed that R^(2) was 0.0955,and RMSE was 3.145.Therefore,we concluded that the gradual linear regression model of vegetation index could accurately estimate the leaf chlorophyll content.The results provide a basis for monitoring the change of chlorophyll contents in vegetation by hyperspectral technology.
作者
何桂芳
吴见
彭建
谷双喜
HE Gui-fang;WU Jian;PENG Jian;GU Shuang-xi(School of Geographic Information and Tourism,Chuzhou University,Chuzhou 239000,Anhui,China)
出处
《西北林学院学报》
CSCD
北大核心
2022年第1期25-32,59,共9页
Journal of Northwest Forestry University
基金
安徽省教育厅高校自然科学研究重点项目(KJ2019A0634)
安徽省教育厅高校自然科学研究重点项目(KJ2018A0434)
安徽省自然科学基金(1808085QC72)。
关键词
叶绿素含量
高光谱
偏最小二乘
逐步线性回归
植被指数
chlorophyll content
hyper-spectral
partial least square
stepwise linear regression
vegetation index
作者简介
第一作者:何桂芳,讲师,硕士,研究方向:3S应用,E-mail:heguifang@chzu.edu.cn。
