Hyperspectral analysis for Magnesium Estimation in Soil

The present study showcases the efficacy of merging conventional soil chemometric analysis with sophisticated hyperspectral imaging methods to precisely evaluate soil nutrients, with a particular emphasis on magnesium concentration. Soil samples were carefully collected from cotton field and tested for magnesium content using titration procedures. The spectral signatures of the soil samples were obtained with a ground-based spectroradiometer and PRISMA hyperspectral photography, which covered a wavelength range of 400 to 2500 nm. Our investigation revealed a significant correlation between magnesium concentration and spectral reflectance in the $765-960 \mathrm{~nm}$ range. Particularly 863 nm wavelength has the highest predictive potential with $\mathbf{R}^{\mathbf{2}}$ of 0.643 and normalized $R M S E$ of 0.001. This wavelength range is useful for constructing accurate remote sensing models for soil nutrient measurement. These findings indicate the potential of hyperspectral data to increase soil analysis from spaceborne sensors, paving the path for better agricultural management methods and long-term soil health monitoring.