一种基于无人机多光谱数据的生物量估算方法研究

A Method of Biomass Estimation Based on UAV Multispectral Data

在多光谱数据中,两数据点间连线的斜率能够大致反映数值的变化规律。为探究斜率能否作为参数估算地上生物量(above ground biomass,AGB),且保持与红边的高线性相关性。本研究使用两个无人机多光谱数据集,利用简单线性回归(simple linear regression,SLR)估算AGB,用决定系数(R2)与均方根误差(root mean squares error,RMSE)来评估线性相关性进而评定其作为估算AGB的能力。结果表明:直接使用斜率进行线性回归时,斜率在不同波长表现不同,在680 nm处的斜率线性相关性最高。在构建的植被指数(vegetation indexes,VIs)中,由红边处斜率参与构建的植被指数同样有着高的线性相关性,并且表现出了比反射率计算的植被指数有更高的线性相关性。

In multispectral data,the slope of the line be⁃tween two data points can roughly reflect the changing regu⁃lation of the data.This article explores whether slope can estimate AGB(Above Ground Biomass)and maintain the high correlation between the red edge and AGB.Drawing on two UAV multi-spectral data sets,this article uses Sim⁃ple linear Regression(SLR)to estimate AGB,and uses R2(coefficient of determination)and RMSE(root mean square error)to evaluate linear correlation,then evaluate its estimate capability of AGB.The results show that when slope is used directly,the slope value behaves differently at different wavelengths,and the linear correlation of the slope at 680 nm is the highest.VIs(vegetation indexes)con⁃structed by the slope at the red edge also has a high linear correlation,achieving a higher linear correlation than that from reflectance.