基于无人机遥感的棉花主要生育时期地上生物量估算及验证

利用棉花主要生育时期的无人机近红外影像数据，提取4种不同的植被指数，通过与棉花地上生物量的实测值建立拟合关系，分析了不同植被指数在棉花各生育时期的估算效果并对其进行了验证。结果表明，随棉花生长，归一化植被指数（NDVI）、宽动态植被指数（WDRVI）、比值植被指数（RVI）和差值植被指数（DVI）均从苗期开始显著增加，其后则表现为基本稳定的“饱和”现象，但棉花实测生物量在不同生育期均有显著差异。植被指数与棉花实测生物量的拟合结果显示：NDVI和DVI的二元线性拟合模型对苗期生物量拟合效果最佳（R²=0.84，RMSE=0.13 kg·m^-2）；WDRVI和DVI的二元线性拟合模型对花蕾期生物量拟合效果最佳（R²=0.87，RMSE=0.52 kg·m^-2）；RVI的非线性拟合模型对花铃期生物量拟合效果最佳（R²=0.79，RMSE=0.95 kg·m^-2）；WDRVI和RVI的二元线性拟合模型对盛铃期生物量的拟合效果最佳（R²=0.86，RMSE=0.96 kg·m^-2）。

Estimation and validation of above-ground biomass of cotton during main growth period using Unmanned Aerial Vehicle (UAV)

Using near-infrared image data collected by Unmanned Aerial Vehicle (UAV) during the main growth period of cotton, four different vegetation indices were extracted to construct optimal estimation model with above-ground biomass (AGB). The results showed that, with growth of cotton, Normalized Difference Vegetation Index (NDVI), Wide Dynamic Range Vegetation Index (WDRVI), Ratio Vegetation Index (RVI), and Difference Vegetation Index (DVI) all increased firstly and then stayed constant. However, AGB of cotton varied significantly among all growth periods. AGB at seedling period was best fitted by binary linear model between NDVI and DVI (R²=0.84, RMSE=0.13 kg·m^-2), while AGB at bud period was best fitted by binary linear model between WDRVI and DVI (R²=0.87, RMSE=0.52 kg·m^-2). At blooming period, AGB was best predicted by nonlinear model of RVI (R²=0.79, RMSE=0.95 kg·m^-2). At boll period, AGB was best estimated by binary linear model between WDRVI and RVI (R²=0.86, RMSE=0.96 kg·m^-2).