High-Spectral Characteristics of Chlorophyll in Phyllostachys edulis at Different Altitudes and Chlorophyll Estima-tion Models
Chlorophyll is an important indicator of vegetation productivity,and the rapid detection of chlorophyll in Phyllostachys edulis is of great significance for the efficient and precise management of agricultural and forestry crops.This study focuses on mountainous P.edulis forests in Shangping Township,Yong'an,Fujian Province,and collects 65 samples of P.edulis leaves from different canopy layers at three altitude gradients(low,middle,and high).Each sample's hyperspectral and chlorophyll mass fraction were measured three times,analyzing the relationship between chlorophyll mass fraction and spectral characteristics along the altitude gradient.The study aims to identify the spectral characteristics and indices that are sensitive to changes in chlorophyll mass fraction.Five parametric regression equations and three machine learning algo-rithms were applied to fit the chlorophyll mass fraction estimation model for P.edulis.The results showed that altitude is a significant factor affecting the spectral characteristics of P.edulis leaves,with chlorophyll mass fraction showing a signifi-cant increasing trend with altitude.The sensitivity of the"three edge"(blue-edged,yellow-edged,and red-edged)param-eters to chlorophyll mass fraction was poor,while both the original spectrum and the first derivative spectrum exhibited sig-nificant multi-band correlations with chlorophyll mass fraction in the visible light range.The sensitive wavelengths for chlo-rophyll were found to be 683 nm and 890 nm in the original spectral curve,and 749 nm in the first derivative spectral curve.The quadratic function regression model established using the red-edge chlorophyll index(ICIred)had an R2 of 0.613.The random forest model(RF)achieved an R2 of 0.852,showing a 38.9％improvement in accuracy compared to the parametric models,with a root mean square error reduction of 61.0％and an average relative error reduction of 38.6％.The ICIred+RF combined model is the best single-variable estimation model for the chlorophyll mass fraction of P.edulis leaves.
High-spectralAltitudinal gradientPhyllostachys edulisChlorophyllSpectral indexMachine learning model
NETL
NSTL
万方数据
