Multispectral-based monitoring of papaya photosynthetic performance using UAV-VTOL for precision agriculture applications

Abstract

Accurate monitoring of crop photosynthetic performance remains a key challenge in precision agriculture, particularly under heterogeneous tropical field conditions where canopy structure, soil background, and plant water status vary within short spatial distances. This study developed a UAV-VTOL-based multispectral monitoring framework for estimating ETR and interpreting plant-level photosynthetic status in papaya. UAV-derived multispectral reflectance and vegetation-index features were paired with field physiological observations at the individual-plant level. Four regression models were compared using repeated cross-validation with hyperparameter tuning and error-based performance metrics. PLSR achieved the most robust performance for ETR estimation, with cross-validated R2 = 0.706, RMSE = 47.08, MAE = 38.60, and nRMSE = 29.63%, followed by ExtraTrees, Random Forest, and SVR-RBF. Although the model captured the dominant ETR gradient, approximately 29% of physiological variability remained unexplained, indicating the need for cautious interpretation and further validation. Nonlinear physiological analysis revealed a threshold-type response of ETR to stomatal conductance, with activation around 0.4–0.5 mol m−2 s−1 and saturation near 220–240 ETR. Interaction surfaces further indicated that high ETR occurred only within constrained stomatal–fluorescence–spectral domains. Explainable modelling identified WDRVI_mean, MSAVI2_mean, and RVI_mean as dominant spectral drivers, highlighting the importance of red–NIR contrast and soil-background correction under open papaya canopy conditions. Productivity-regime analysis classified plants into stress, transitional, and optimal photosynthetic states. These findings indicate that UAV-VTOL multispectral sensing can provide a preliminary plant-level diagnostic framework for assessing papaya photosynthetic performance, while broader validation across seasons, cultivars, and environmental conditions is required before operational deployment.