Automated construction of levels of detail models for grotto facades

In the 3D digital preservation of grotto temples, complex grotto facades require many faces, resulting in large datasets and high computational costs that limit usability across application scenarios. Currently, the Levels of Detail (LoDs) granularity for grotto facades is unclear, and automated generation methods are lacking. To address this, four LoDs are defined for grotto facades, representing sculptural details, spatial structure, spatial massing, and horizontal 2D representation, with automated construction methods proposed for each LoD. Finally, taking Cave 38 of the UNESCO World Heritage Yungang Grottoes as an example, four LoD models were constructed, and geometric deviation, data volume, and modeling efficiency were analyzed. The results show that the proposed method reduced LoD3–LoD0 generation time from 249.1 min required by manual modeling to 59.44 min, while LoD2 reduced faces by over 99% compared with the original mesh, reducing manual processing and modeling time and improving usability across scenarios.