Mandala3D: A Configurable Pipeline for Processing Textured 3D Scans in Cultural Heritage Documentation

The digitization of cultural heritage assets using 3D scanning technologies offers unprecedented opportunities for preservation, analysis, and dissemination. However, raw 3D scan data often requires significant processing to become truly usable, involving steps like format conversion, cleaning, texturing, and metadata enrichment. Manual execution of these steps is time-consuming, error-prone, and lacks consistency, hindering large-scale digitization efforts. This manuscript introduces Mandala3D, a configurable, open-source Python pipeline designed to automate the conversion of textured 3D mesh models (primarily OBJ and PLY formats) into accurately colored, analysis-ready point clouds (PLY and LAZ formats). Mandala3D integrates functionalities from established libraries like Open3D and Trimesh, providing a cohesive workflow that includes mesh validation, optional pre-processing (healing, decimation), robust texture and vertex color application using barycentric interpolation, point cloud post-processing (outlier filtering, normal estimation), comprehensive metadata generation (including checksums and processing provenance), thumbnail creation, and automated reporting. By offering a modular, configurable, and automated approach, Mandala3D aims to enhance the efficiency, consistency, and data quality associated with processing 3D scans in cultural heritage documentation, thereby supporting more effective digital preservation and research.