The feasibility of using a low-cost near-infrared, sensitive, consumer-grade digital camera mounted on a commercial UAV to assess Bambara groundnut yield

Accurate, timely, and non-destructive early crop yield prediction at the field scale is essential in addressing changing production challenges mitigating impacts of climate variability. Unmanned aerial vehicles (UAVs) are increasingly popular recent years for agricultural remote sensing applications such as forecasting precision agriculture (PA). The objective this study was to evaluate performance a low-cost UAV-based technology Bambara groundnut prediction. A multirotor UAV equipped with near-infrared sensitive consumer-grade digital camera used collect image data during 2018 growing season (April August). Flight missions were carried out six times critical phenological stages life-cycle monitored crop. Yield recorded harvest. Four vegetation indices (VIs) namely normalized difference index (NDVI), enhanced 2 (EVI2), green (GNDVI), simple ratio (SR) generated from Red-Green-Near Infrared bands calculated using georeferenced orthomosaic images. Pearson’s product-moment correlation coefficient (r) Bland–Altman testing showed significant agreement between remotely proximally sensed VIs. Significant positive correlations found four VIs yield, strongest relationship observed SR podfilling stage (r = 0.81, P < 0.01). Multi-temporal accumulative improved significantly best being ∑SR interval maturity 0.88, accumulated resulted higher accuracy determination (R2) 0.71, root mean square error (RMSE) 0.20 absolute percentage (MAPE) 14.2% than spectral single (R2 0.68, RMSE 0.24, MAPE 15.1%). Finally, map model developed, better understand within-field spatial variations future site-specific or variable-rate application operations.