Extraction and restoration of scratched murals based on hyperspectral imaging—a case study of murals in the East Wall of the sixth grotto of Yungang Grottoes, Datong, China

A novel method using hyperspectral imaging significantly enhances the accuracy of extracting and restoring scratched murals, addressing critical challenges in cultural heritage preservation. The study introduces a technique combining Principal Component Analysis (PCA) with a high-pass filter to improve image information, followed by multi-scale bottom hat transformation and Otsu threshold segmentation for precise deterioration extraction. Morphological transformations and connected component analysis further refine the results, while an improved exemplar-based region filling method restores the scratched areas effectively. The approach outperforms other methods in both extraction accuracy and structural integrity preservation, offering a promising solution for virtual restoration of deteriorating murals at UNESCO World Heritage Sites like Yungang Grottoes.

Qiao, K., Hou, M., Lyu, S., & Li, L. (2024). Extraction and restoration of scratched murals based on hyperspectral imaging—a case study of murals in the East Wall of the sixth grotto of Yungang Grottoes, Datong, China. Heritage Science, 12(1). https://doi.org/10.1186/s40494-024-01215-3