Abstract This study presents a novel non-invasive dual wavelength photoacoustic imaging (DWPAI) system developed to detect distinct spectral signatures of historical ink families with high sensitivity. The recorded images were processed through unsupervised machine learning algorithms, to provide a quantitative indicator of the ink type based on the ratio of photoacoustic signals following excitation in the visible (532 nm) and near-infrared (1064 nm). The performance of the system was demonstrated using three inks commonly employed in 19th-century Europe and Greece, specifically Chinese ink, sepia ink, and iron gall ink. Chinese ink provided signal ratios around 0.9, sepia ink yielded consistent ratios at 0.6, whereas iron gall ink exhibited considerably higher ratios in the order of 1.7. Our findings demonstrate the ability of DWPAI to reliably and non-destructively discriminate between major ink families independent of the writer or the nib type used. Apart from guiding conservators in safe and effective restoration interventions, DWPAI may provide archaeologists and art historians with complementary insights into the dating, origin, and material characterization of various types of artifacts. Its demonstrated accuracy and versatility establish DWPAI as a powerful diagnostic tool for cultural heritage research, archaeology, and conservation science.
Pirgianaki et al. (Thu,) studied this question.