Publication: ''UAV multispectral to hyperspectral reconstruction based on deep learning and radiative transfer models for crop nitrogen monitoring''

The study introduces M2H–SWIR, a novel multispectral-to-hyperspectral reconstruction framework that integrates the physically based PROSAIL-PRO radiative transfer model with deep learning. The approach enables the reconstruction of continuous spectral reflectance across a broad wavelength range of 400–2500 nm, extending beyond the visible and near-infrared bands captured by conventional UAV multispectral cameras to include the short-wave infrared (SWIR) region.

This is particularly relevant for precision agriculture, as SWIR wavelengths contain important spectral information related to crop biochemical and structural properties, including features associated with nitrogen status. While reconstructing the SWIR region from multispectral observations remains challenging due to factors such as water absorption and canopy structural effects, the results highlight the strong potential of combining UAV remote sensing, radiative transfer modeling, and artificial intelligence to extract richer spectral information from comparatively affordable multispectral sensors.

Publication: Wang, J., Belwalkar, A., Meyer, S. T., Li, F., Herrmann, I., & Yu, K. (2026). UAV multispectral to hyperspectral reconstruction based on deep learning and radiative transfer models for crop nitrogen monitoring. International Journal of Applied Earth Observation and Geoinformation, 150, 105364.

DOI: 10.1016/j.jag.2026.105364