Using structure-from-motion photogrammetry, our surveyors capture 40 to 80 overlapping photographs of each gravestone from multiple angles and elevations. These images are processed through specialized software (Agisoft Metashape or Reality Capture) to generate precise 3D mesh models. The resulting models preserve the exact geometry of the stone — including deterioration patterns — in a permanent digital form that can be studied long after the physical marker has degraded.

Each grave is georeferenced in the field using the best location source available at capture time. The current mobile workflow records handset GPS with averaging and exposes the reported meter-level accuracy to the surveyor. When teams collect external GNSS receiver data or raw observation logs, those files can be retained for later PPK or receiver-based correction workflows. All accepted coordinates are stored as PostGIS spatial data for mapping, distance calculations, and sector analysis.

Gravestone inscriptions are among the most important historical data we capture. Our system uses Google Gemini Vision to automatically transcribe text from photographs, extracting names, birth and death dates, and biographical details. The AI output is always reviewed by a human researcher before being accepted into the archive. For worn or damaged stones, AI-enhanced image processing (contrast boost, infrared simulation) improves legibility.

Beyond individual marker documentation, we are building a sector-level digital twin of the entire cemetery using drone-captured orthomosaic imagery. Our drone operators fly systematic grid patterns at low altitude, generating high-resolution top-down imagery that is stitched into a georeferenced orthomosaic. This enables cemetery-wide analysis of marker density, condition distribution, and preservation priorities. The long-term vision is a fully interactive 3D model of the entire Concordia grounds.

All data collected is stored in a production PostgreSQL database with PostGIS spatial extensions, hosted on Supabase with automated backups and row-level security. The public archive is designed to be permanent and interoperable — records can be exported in standard formats for use with genealogy platforms, academic databases, and cultural heritage systems. Our schema is fully documented and open-source, and we plan to publish a full data export under Creative Commons license.