First Application of UAV-Based Hyperspectral Imaging (HAIP Camera on DJI M300) for Monitoring Oat Hybrids and Chemical Treatments Under Climate Change Conditions.

Abstract

UAV-based hyperspectral imaging was used to monitor oat hybrids treated with silicon, Sulphur, and their combination, capturing high-resolution spectral data at key growth stages with an HAIP camera mounted on a DJI Matrice 300 drone. Machine learning methods, including Random forests, Support vector machines, Gradient boosting, and CatBoost, were applied to analyze the data, with CatBoost achieving the highest accuracy in classifying treatments. The research successfully distinguished between oat hybrids and detected their physiological responses to chemical treatments, with Minimum Noise Fraction combined with Principal Component Analysis proving most effective for feature extraction. Overall, the research establishes UAV hyperspectral imaging as a powerful, non-destructive precision agriculture tool for supporting climate-resilient cereal production and enabling rapid crop management decisions.