Predictive Modeling of Soil Chromium Content Using Hyperspectral Spectroscopy

This study explores the quantification of chromium heavy metal compounds (CrO3 and Cr2O3) in soil using VNIR-SWIR spectroscopy and employs partial least squares regression (PLSR). Significant results were obtained, with R2 values ranging from 0.72- 0.86 for CrO3 and 0.86 - 0.93 for Cr2O3. Specific regions of the electromagnetic spectrum were identified as being most important in terms of chromium spectral features, offering insights into the link between spectral characteristics and heavy metal concentrations. Given that chromium concentrations in agricultural soils can reach dangerous levels, this study's findings emphasize the necessity of precisely determining chromium levels in agricultural soils.