Spatial distribution analysis of surface moisture content in ancient murals and polychrome sculptures based on hyperspectral imaging

Surface moisture content is a key driver of deterioration in murals and polychrome sculptures, yet non-destructive macroscopic mapping methods remain limited. Taking the murals and polychrome sculptures of Beijing Dahui Temple as a case study, this study combined hyperspectral imaging (HSI) with fractional-order derivative (FOD) analysis to enhance moisture-related spectral features. Partial least squares regression (PLSR) and support vector regression (SVR) models were developed to estimate and map surface moisture content. FOD preprocessing significantly improved the correlation between spectral data and measured moisture. The results show that the murals were generally dry and stable, with moisture contents of 3.79%–5.09%, whereas the polychrome sculptures were wetter, ranging from 7.09% to 8.55%, with local areas reaching 10%–11%. This study demonstrates a feasible non-destructive approach for quantitative mapping and spatial visualization of surface moisture variations in murals and polychrome sculptures.