Waterlogged archaeological wood samples may degrade during long-term immersion in microbial-activity environments, which causes its biodegradation. Simultaneous dynamic vapor sorption (SDVS) and two-dimensional correlation infrared (2D COS-IR) spectroscopy reveal the degradation inhomogeneity of waterlogged fir wood from the Shengbeiyu shipwreck. The waterlogged and reference wood exhibit type II sorption isotherms. The equilibrium moisture contents of waterlogged archaeological fir wood from a decay region (WFD) were 22.5% higher than those of waterlogged archaeological fir wood from a sound region (WFS). WFD exhibits a higher measurable sorption hysteresis than WFS, implying greater variation in the surface moisture content in the WFD region compared to the WFS region, which may compromise the dimensional stability of the shipwreck. 2D COS-IR spectra confirmed the inhomogeneous degradation of the waterlogged wood via numerous mechanisms. The efficacy of SDVS and 2D COS-IR spectroscopy in the evaluation of the degradation state of waterlogged wood was demonstrated. This study verifies the existence of hygroscopic and chemical differences between visually similar samples from the same shipwreck.

The Nanhai No. 1 shipwreck is an ancient wooden ship in the Southern Song Dynasty. Currently, serious challenges of microbial diseases exist on the hull wood. This study aimed to obtain microbial samples from the ship hull in December 2021 and analyze the microbial diseases through scanning electron microscopy and high-throughput sequencing to preserve the Nanhai No. 1 shipwreck. The biodegradation mechanism of diseased microorganisms was explored through whole genome sequencing and the detection of enzyme activity and gene expression levels of diseased microorganisms under different conditions. The results showed that there was obvious fungal colonization on the surface of the hull wood and Fusarium solani NK-NH1 was the dominant disease fungus on the surface. NK-NH1 has strong cellulose and lignin degradation ability. Its whole genome size is 52,389,955 bp, and it contains 17,402 genes. It has a variety of key enzyme genes involved in cellulose and lignin degradation. The NK-NH1 dominant degrading enzyme lignin peroxidase has the highest enzyme activity at pH = 4, NaCl concentration of 30%, and FeSO₄ concentration of 50 mg/L, while laccase has the highest enzyme activity at pH = 4, NaCl concentration of 10%, and FeSO₄ concentration of 100 mg/L. The above research results prove that NK-NH1 is a key fungus to the biodegradation of ship hull wood when it is exposed to air, low pH, high salt, and rich in sulfur iron compounds. This study provides a theoretical basis for the preservation of the Nanhai No. 1 shipwreck.

This study investigates the classification of pigment-manufacturing processes using deep learning to identify the optimal model for cultural property preservation science. Four convolutional neural networks (CNNs) (i.e., AlexNet, GoogLeNet, ResNet, and VGG) and one vision transformer (ViT) were compared on micrograph datasets of various pigments. Classification performance indicators, receiver-operating characteristic curves, precision–recall curves, and interpretability served as the primary evaluation measures. The CNNs achieved accuracies of 97–99%, while the ViT reached 100%, emerging as the best-performing model. These findings indicate that the ViT has potential for recognizing complex patterns and correctly processing data. However, interpretability using guided backpropagation approaches revealed limitations in the ViT ability to generate class activation maps, making it challenging to understand its internal behavior through this technique. Conversely, CNNs provided more detailed interpretations, offering valuable insights into the learned feature maps and hierarchical data processing. Despite its interpretability challenges, the ViT outperformed the CNNs across all evaluation metrics. This study underscores the potential of deep learning in classifying pigment manufacturing processes and contributes to cultural property conservation science by strengthening its scientific foundation for the conservation and restoration of historical artifacts.

In 2013, several tombs were discovered and excavated in Southeastern Beijing, China, yielding various burial products. Among these were pottery towers, a representative artifact of the Han Dynasty rarely found in the area. Many studies on architectural aspects, such as construction type and construction situation, have been conducted based on the excavated pottery towers, but only a few have examined their raw materials or pigments. In this study, black, white, and red pigments were identified as carbon black, calcite, and cinnabar, respectively, through a multi-analysis approach. The manufacturing method of the pigment was established based on the crystal form and particle size of the pigment, by factoring in the pigment characteristics, which cannot be distinguished by component analysis and crystal structure analysis. We recommend that a continuous database be prepared and used in the future, not only for an interpretation of ancient pigments but also to identify the factors to be considered (physical characteristics, such as pigment particle size among others) when estimating the manufacturing process and conservation treatment.

The development of convolutional neural network (CNN) techniques has advanced image classification in various fields, including autonomous vehicles, medical care, and cultural heritage science^1,2,3. This study employed CNN techniques to process large numbers of micrographs, build a dataset of pigment microstructures, and classify their manufacturing processed. As research that automatically classified such micro-morphological features using CNNs remains limited, this study provides an early framework demonstrating the potential of deep learning-based morphological analysis in pigment research within cultural heritage. A dataset of 2654 SEM micrographs was established and analyzed using four architectures—AlexNet, GoogLeNet, ResNet, and VGG. Evaluation metrics such as accuracy, precision, recall, F1-score, and confusion matrix were applied, with VGG16 achieving the highest overall performance. The CNN models reached over 97% accuracy using an 80/20 split for cross-validation. These results indicate that the manufacturing process of a pigment can be inferred rapidly by comparing and classifying its micrographs.