Automated Digital Documentation and Interpretation of Subsurface Cultural Heritage Based on 3D Geometry and AI Inference

Abstract. This study presents a digital documentation and interpretation methodology for subsurface cultural heritage artifacts based on 3D geometry acquisition and AI inference. A digital solution was developed to automate the conventional manual drawing process of archaeological artifacts, significantly reducing the required time from approximately three hours per artifact to about 23 minutes. Using high-precision 3D scanning and standardized data preprocessing, external contours and cross-sections were automatically extracted, while subjective lines were inferred based on curvature analysis and AI-driven interpretation of accumulated surveyor data. Additionally, visualization data replicating surface patterns and textures were generated through curvature-based shader processing. The proposed approach enhances the objectivity and reliability of artifact documentation by minimizing human error and improving data consistency. It also offers new possibilities for artifact analysis, including the identification of artifacts produced by the same maker through pattern recognition techniques. The results suggest that the developed system can significantly advance digital workflows in archaeological research and heritage conservation, providing a scalable framework for future applications.