Gustav In April 2019, Notre-Dame de Paris caught fire. The spire collapsed in under an hour. Centuries of stone carving, gilded altarpieces, and structural detail that no living architect fully understood began to disappear in real time. What made the subsequent restoration effort possible - and far more accurate than it would have been a decade earlier - was that a historian named Andrew Tallon had spent years producing a comprehensive laser scan of the entire cathedral. That data became the ground truth for reconstruction.
It is one of the clearest recent examples of what digital preservation actually means in practice. Not a backup of a file. Not a photograph. A complete geometric record, accurate to a few millimetres, of an object that no longer exists in its original form.
What photogrammetry actually is
Photogrammetry is the process of deriving three-dimensional geometry from two-dimensional photographs. You capture an object from dozens or hundreds of overlapping angles, feed the images into processing software, and the algorithm reconstructs the geometry by identifying matching points across frames and calculating their position in 3D space.
The output is a dense point cloud - millions of individual spatial coordinates - which is then processed into a mesh and draped with a photographic texture derived from the source images. Done well, the result is a model accurate enough to measure, analyse, and reproduce. The object is not just documented - it is captured.
The two tools I use for this work are Agisoft Metashape and RealityCapture. Metashape has long been the standard in heritage and scientific applications - it is reliable, well-documented, and widely trusted in academic contexts. RealityCapture (developed by Epic Games) processes at significantly higher speed, handles very large datasets efficiently, and produces excellent results - it is increasingly used in film, games, and large-scale heritage projects where processing time matters. For high-quality statue and sculpture work, both tools produce results that hold up to professional scrutiny.
The camera matters more than people think
Photogrammetry software has become remarkably accessible - but the quality of the input data determines the quality of the output. A high-resolution DSLR with a sharp prime lens, shooting in controlled or consistent lighting, will produce point clouds with far more detail than smartphone photography.
For sculpture and statue work specifically, the geometry of the subject creates challenges. Deep undercuts, sharp crevices, highly polished or reflective surfaces - all of these are difficult for photogrammetry to handle cleanly. Managing them requires thinking carefully about capture angle coverage, using additional lighting where needed, and sometimes applying a temporary matte spray to reduce specular reflections on particularly problematic surfaces.
The capture session is where the quality is established. Processing software cannot recover detail that was not captured.
The physical object will eventually change - through damage, decay, restoration, or loss. The 3D record does not. Every scan done today is an insurance policy against a future we cannot predict.
What you can do with a preservation scan
A high-quality photogrammetry scan of a historical object is not a single-use asset. The same data can serve multiple purposes simultaneously.
For academic research: measurements, proportional analysis, surface texture documentation, comparison across time if multiple scans exist. For exhibition: interactive 3D viewing, digital reproductions, VR walkthroughs that let visitors engage with objects that are too fragile to handle or physically inaccessible. For restoration: if part of an object is damaged or lost, the scan provides the ground truth for what was there. For archive: an institution that holds the scan holds the object, in the sense that matters most - its exact geometry and surface detail - even if the physical item is destroyed or dispersed.
Museums and heritage organisations are increasingly recognising this. Scanning collections is moving from a specialised research activity to a standard part of collection management.
The case for doing it now
The argument for scanning is strongest for objects that are already showing deterioration - outdoor sculptures exposed to weather, stone surfaces affected by pollution, wooden or organic objects in unstable storage environments. The scan captures the current state; every year that passes without scanning is a year of degradation that cannot be recovered.
But the argument applies to objects in good condition too. "Good condition now" does not mean "good condition indefinitely." Objects are moved, loaned, damaged in transit, affected by changes in temperature and humidity. A scan made today under stable conditions is more useful than a scan made after something goes wrong.
The cost of a scan is small relative to the cost of restoration, insurance, or loss. The data lasts as long as the storage format it lives in - and file formats are far easier to migrate than physical objects are to restore.
What the process looks like in practice
For the historical statue preservation work I have done, the workflow starts with site assessment - understanding the lighting conditions, the geometry of the subject, the access constraints, and what level of detail the end use case requires. A scan intended for academic measurement has different requirements than one intended for a real-time VR exhibition display.
Capture typically involves several hundred photographs taken from systematically varied angles, with consistent overlap. Post-processing in RealityCapture and Metashape handles alignment, dense point cloud reconstruction, mesh generation, and texture baking. The cleaned mesh is then prepared for whatever the downstream use requires - whether that is a web viewer, a high-polygon archive master, or a reduced-polygon version suitable for real-time rendering.
The result is a dataset that represents the object more completely than any photograph, drawing, or written description ever could.
Who this is for
Museums and cultural institutions with objects that need formal documentation. Archaeologists and researchers who need accurate geometric data. Local heritage organisations that want to create digital versions of local landmarks or sculptures. Architects working on restoration projects. Private collectors who want a permanent record of significant pieces.
The technology is no longer experimental or expensive relative to what it produces. The barrier is mostly awareness - not everyone knows that a scan of this quality is achievable with the right equipment and process, or that the resulting data can serve the range of purposes it can.