Combined use of agro-climatic and very high-resolution remote sensing information for crop monitoring

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.