Multispectral Imaging and Spectral Unmixing for Characterization of Relative Phase Composition of Steel Corrosion Products

The integrity of steel structures is significantly impacted by corrosion, necessitating precise assessment methods. This study introduces an innovative approach using multispectral imaging (MSI) and spectral unmixing (SU) to analyze the relative composition of corrosion phases such as Lepidocrocite, Magnetite, Goethite, and Hematite on steel samples. We developed a specialized MSI setup covering the visible to near-infrared (NIR) spectrum and evaluated various linear and nonlinear SU techniques. Our findings reveal the high accuracy and efficiency of the MSI and SU systems in determining corrosion phase distribution and composition. This method was rigorously validated against X-ray diffraction (XRD) and Rietveld analyses, affirming its potential as a nondestructive, widespan tool for corrosion assessment. These results have significant implications for improving the longevity and safety of steel structures through more precise corrosion monitoring.