Artificial intelligence assists the identification and application of rural heritage wall surface damage in Zhejiang
Published in Social Sciences, Mathematical & Computational Engineering Applications, and Arts & Humanities
Our paper addresses the pressing issue of preserving rural cultural heritage in Zhejiang Province, where traditional dwelling walls face damage due to humid conditions. Utilizing the YOLOv8 deep learning model, we enhance automated damage detection through multi-layer feature extraction and feature fusion. Field experiments in Hangzhou and Lishui demonstrated high accuracy in detecting yellowing (0.8) and stains (F1-score 0.71), especially under high-resolution conditions. Additionally, a Raspberry Pi 5-based mobile detection system offers real-time monitoring, providing a cost-effective solution for cultural heritage protection. Read the full paper here. https://www.nature.com/articles/s40494-025-01725-8