Water bodies are critical to the environment, providing numerous ecological benefits; however, human activities increasingly threaten their quality. Natural water systems exhibit regional variability, dominated by organic and inorganic species, rendering in-situ measurements insufficient. Current remote sensing methods often overlook the impact of surface light components, which vary with solar radiation and wave intensity. This study demonstrates an approach for water quality monitoring utilizing spectral reflectance and polarization in the visible and near-infrared regions. A line spectrometer with a polarization filter was employed for hyperspectral measurements under simulated wave conditions. Chlorophyll (Chl) and suspended sediment (SS) pollution were simulated using locally sourced products at various concentrations. The contaminant reflectance was computed, and the polarization components were analyzed using the Pickering method and Stokes vectors. Normalization and continuum removal techniques ensured reliable comparisons across the spectra. The spectral angle mapper (SAM) algorithm quantified the similarity between unpolarized wave conditions and total reflectance spectra under calm conditions, while spectral entropy quantified wave effects compared to calm water. The results indicated that the polarized components of Chl and SS reflectance were minimal in calm water but increased under wave conditions, particularly at lower wavelengths. Higher contaminant concentrations exhibited greater spectral similarity, with lower SAM values indicating reduced specular reflections. The raw and normalized unpolarized reflectance spectra displayed characteristic features; however, the reflectance at high concentrations was lower than anticipated, likely due to spectrometer sensitivity and strong water absorption. The degree of Linear Polarization (DoLP) analysis revealed distinct scattering behaviors: Chl exhibited a lower DoLP than SS. Wave conditions enhanced the DoLP due to increased specular reflections. Overall, the method effectively separates reflection components but is sensitive to measurement conditions, emphasizing the necessity to account for angles and water conditions when estimating contaminants.
Read full abstract