Accurate and real-time yield forecasting is one of the main pillars for decision making in farming and thus for farmers’ profitability. Biomass has been traditionally predicted by multi- and hyperspectral vegetation indices from low- and medium-resolution platforms. This research work aimed to assess the accuracy of the combined use of agro-climatic information and very high-resolution products obtained with RGB cameras mounted on unmanned aerial vehicles (UAVs) for biomass predictions in maize (Zea mays L.). Two agro-climatic predictors, reference evapotranspiration (ETo) and growing degree days (GDDs), and twelve vegetation indices (VIs) derived from RGB bands were calculated for the entire growing cycle. The root mean squared error (RMSE) of the model that considers only GDD to estimate total dry biomass (TDB) was 692.7 g m⁻², which was reduced to 509.3 g m⁻² when introducing as predictor variables the VARI and GLI vegetation indices. Difficulties in the radiometric calibration of consumer grade RGB cameras together with sources of error such as the bidirectional reflectance distribution function and the blending algorithms in the photogrammetry processing could decrease the applicability of the obtained relationship and should be further evaluated. This study illustrated the advantage of the combined use of agro-climatic predictors (GDD) and green-based VIs derived from RGB consumer grade cameras for biomass predictions.

准确和实时的产量预测是农业决策以及农民盈利能力的主要支柱之一。传统上，通过低分辨率和中分辨率平台的多光谱和高光谱植被指数来预测生物量。这项研究工作旨在评估将农业气候信息和超高分辨率产品组合使用的准确性，该产品是通过安装在无人机（UAV）上的RGB摄像机获得的，用于预测玉米中的生物量（玉米）L.）。计算了两个农业气候预测因子，即参考蒸散量（ETo）和生长度日（GDDs），以及从RGB波段得出的十二个植被指数（VIs），用于整个生长周期。仅考虑GDD估算总干生物量（TDB）的模型的均方根误差（RMSE）为692.7 g m ^-2，已降至509.3 g m ^-2当引入VARI和GLI植被指数作为预测变量时。消费级RGB相机的辐射校准的困难以及诸如摄影测量处理过程中的双向反射率分布函数和混合算法之类的误差源可能会降低所获得关系的适用性，因此应进一步评估。这项研究说明了结合使用农业气候预测器（GDD）和基于RGB消费级相机的绿色VI进行生物量预测的优势。