Airborne hyperspectral imaging of nitrogen deficiency on crop traits and yield of maize by machine learning and radiative transfer modeling

•Photosynthetic capacity of maize was predicted from airborne hyperspectral imagery.•Radiative transfer modeling surrogates and PLSR were used to predict leaf and canopy traits.•Soil and leaf angle information can reduce ill-posed radiative transfer model retrieval.•Canopy structure spectral signals enhance the accuracy of predicting canopy photosynthetic traits.•Synergistic use of process-based and data-driven approaches facilitates trait retrieval.