论文部分内容阅读
在BRDF测试系统环境下利用ASD便携式野外光谱仪采集樟树叶片光谱,并用UV2450-紫外可见分光光度计对观测叶片进行叶绿素含量测定.考虑到植物色素(叶绿素和类胡萝卜素)对叶片反射光谱的影响主要体现在可见光波段,选取400~900 nm范围波段光谱反射率与叶片叶绿素含量反演偏最小二乘法(PLS)模型,其中29个样本用于建模,10个样本用于验证,结果表明:当主成分个数为4时,PLS模型具有最佳的效果,4个主成分累计解释了99.91%的自变量信息和89.71%的因变量信息,此外,PLS模型能够充分利用高光谱信息,具有较高的精度和稳定性.通过与原始光谱和一阶导数光谱拟合的估测模型进行对比分析,得出PLS模型无论是从建模样本精度还是验证的误差方面均优于这两种传统的模型,适合于利用高光谱数据进行叶绿素含量的估测.
The spectrum of camphor tree leaves was collected by using ASD portable field spectrometer in the environment of BRDF test system, and the chlorophyll content of the leaves was measured by UV2450-UV-Vis spectrophotometer.Considering the effect of plant pigments (chlorophyll and carotenoid) on the reflectance spectra of leaves, In the visible wavelength range, PLS model was used to reflect spectral reflectance in the range of 400-900 nm and leaf chlorophyll content. 29 samples were used for modeling and 10 samples were used for validation. The results showed that when the main PLS model has the best effect when the number of components is 4, 99.91% of the independent variable information and 89.71% of the dependent variable information are explained by the four principal components. In addition, the PLS model can make full use of the hyperspectral information and has a higher The accuracy and stability of PLS model are compared with those of the original model and the first derivative model.The results show that the PLS model is superior to the two traditional models , Suitable for estimating chlorophyll content using hyperspectral data.