Physics-guided AI for robust crop traits estimation

Conducted at the University of Nebraska-Lincoln

Overview

To improve crop nutrient estimation, we proposed a novel hybrid method that integrates the computational efficiency of AI models with the robustness of physics-based Radiative Transfer Models (i.e., physics-guided AI). This new approach not only achieved higher accuracy than the conventional methods but also demonstrated better robustness across diverse field locations, various crop growth stages, and different growing seasons. With its demonstrated success, this method holds significant promise for application to other common crop species or traits.

Publications

Li, J., Ge, Y., Puntel, L. A., Heeren, D. M., Bai, G., Balboa, G. R., … & Shi, Y. (2024). Integrating UAV hyperspectral data and radiative transfer model simulation to quantitatively estimate maize leaf and canopy nitrogen content. International Journal of Applied Earth Observation and Geoinformation, 129, 103817. paper link

Li, J., Wijewardane, N. K., Ge, Y., & Shi, Y. (2023). Improved chlorophyll and water content estimations at leaf level with a hybrid radiative transfer and machine learning model. Computers and Electronics in Agriculture, 206, 107669. paper link