Author(s)

Yixuan Chu

Corresponding Author

Yixuan Chu

Abstract

The preservation of Sichuan’s watchtowers, significant cultural heritage sites, is increasingly threatened by environmental factors, urbanization, and inadequate conservation efforts. This study proposes an integrated framework combining environmental monitoring and machine learning. Sensor networks tracked factors such as temperature, humidity, and air quality, while drone-based photogrammetry captured high-resolution spatial data for structural analysis. Machine learning models, including CNNs with 92% accuracy for crack detection and RNNs for forecasting deterioration trends, provided actionable insights for repair prioritization. A Decision Support System (DSS) integrated real-time data with predictive outputs, optimizing maintenance planning. This scalable framework enhances the sustainable preservation of Sichuan’s watchtowers and offers a replicable solution for global heritage conservation.

Keywords

Sichuan Watchtowers, Machine Learning, Ecological Monitoring, Structural Integrity