A Digital Intelligence System for 3D Reconstruction and Cultural Feature Analysis of Ancient Window Paper-Cuts in Southern Hunan

Facing the challenge of irreversible loss of artistic details caused by the complexity of 3D patterns and limitations of manual surveying, a full-chain technical system is proposed, integrating 3D laser scanning, oblique photogrammetry, multi-source data fusion, and deep learning algorithms. This method enables millimeter-level three-dimensional data acquisition, parametric model reconstruction, and quantitative analysis of decorative features. Experimental results show that the proposed system achieves an average geometric error of 0.12 mm in three-dimensional reconstruction, a pattern restoration completion rate of 98.6%, and a regional style classification accuracy of up to 95.5%. This study not only establishes a high-precision digital protection paradigm for endangered architectural decorations but also offers a cross-disciplinary approach for integrating architectural history, art archaeology, and digital humanities.