Non-invasive pigment analysis of the Kizil Grotto murals by hyperspectral imaging

Abstract As an important Buddhist site on the Silk Road, the Kizil Grottoes in Xinjiang have long faced degradation such as weathering and fading. Non-invasive pigment analysis is essential for scientific study and conservation of cultural heritage. This study focused on a representative mural section in Cave 171 of the Kizil Grottoes. Based on a constructed pigment hyperspectral database, the spectral characteristics of representative pigments were analyzed. Visible and near-infrared point-based reflectance spectra and hyperspectral imaging data were acquired, with X-ray fluorescence spectroscopy used for verification. Multiple Endmember Spectral Mixture Analysis (MESMA) was applied for pigment identification and spatial distribution mapping. Results show that different pigments exhibit distinguishable spectral features, and MESMA can utilize subtle spectral differences for spectral mixture analysis. This study demonstrates the potential of hyperspectral data processing methods for non-invasive pigment identification and spatial characterization of mural surfaces, providing a reference for the analysis of cultural heritage paintings.