Heritage 3D digitisation 3D Gaussian Splatting

Digitising a historic monument: the 6 steps of a successful 3D capture

By Olivier Kerremans ·

Digitising a historic monument in 3D has become accessible, economically viable and technically accomplished. The heritage institutions (museums, castles, cathedrals, foundations) that take an interest in it often do not know exactly where to start or what to expect from the process. This article describes the 6 steps of a successful 3D capture of a monument, from initial scouting to long-term maintenance, with concrete lessons from a Gothic cathedral project.

For the general technology context, see our complete guide to 3D Gaussian Splatting and our comparison of digital twin technologies.


Why digitise a historic monument in 3D

Four objectives drive these projects. Identifying the right one from the start changes the technical specification.

Archiving before restoration

A restoration project changes the state of a monument. Documenting the initial state in photorealistic 3D creates an opposable reference, useful for comparing before and after, and for passing on to future generations a state that the works will make disappear.

Augmented cultural mediation

A physical visit remains irreplaceable. But it can be enriched. An immersive path viewable beforehand prepares the visitor, broadens the audience to distant publics, and creates multilingual entry points for international tourism.

Accessibility

Not everyone can physically access a monument. People with reduced mobility, sick children, geographically distant people, schoolchildren without a travel budget. A quality virtual tour opens access without degrading the experience.

Digital outreach and tourism

A 3D presence on the institutional website, on tourism platforms, or in partnership with cultural media generates traffic, shares, and physical-visit conversions. It has become a significant marketing lever for sites that depend on tourist flow.


Step 1, qualify the need

The first classic mistake, rushing into the capture without having clarified the final use. What you capture and how you deliver it depend on the purpose.

Three questions to settle first.

Archive or public experience? An archive prioritises maximum fidelity and long-term conservation of the source file. A public experience prioritises smoothness, editorial, and compatibility with your website and channels.

Target audience? French-speaking local visitors, international tourists, schoolchildren, researchers, patrons. Each audience calls for a different narration, a language, a level of historical depth.

Distribution channels? Institutional website only, multi-platform integration, dedicated app, on-site kiosk, presentation at an event. The channel constrains the delivery format.

Our Heritage sector page details the common uses by type of institution.


Step 2, technical scouting

Once the need is qualified, the site is scouted with the provider. This preparatory visit avoids most capture problems.

Volume of the location

A chapel, a church, a cathedral, a museum with several rooms, a castle with grounds. The volume determines the capture time, the number of captures, and the logistics.

Lighting conditions

Natural light changes over the course of the day. On a monument with stained glass, lighting is central. We often plan the capture for a stable light slot (morning or late afternoon depending on the orientation).

Attendance and possible slots

Capture with the public, without the public, partly with the public. Each option has its advantages. A slot without the public gives a purer result. A capture during opening hours shows the place alive, which can be an editorial choice.

Heritage constraints

Specific authorisations (heritage services, curator, owner), restrictions on certain areas (fragile works, sacred spaces), access conditions (scaffolding, heights, narrow stairs).

Presence of elements under restoration

If the monument is under works, you have to decide whether to capture the “under works” state as a document of the process, or wait for completion to capture a restored state.


Step 3, the capture

A few hours on site, sometimes spread over two days for a large monument.

Typical duration by volume

Type of siteCapture duration
Chapel, small museum1 to 3 hours
Medium-sized church3 to 6 hours
Cathedral6 to 10 hours, often over 2 days
Castle with grounds1 to 3 days
Multi-room museum1 to 2 days

Equipment used

For 3DGS capture, we adapt the equipment to the project. A specialised ground, aerial or combined camera depending on the volume, the height and the access constraints specific to the monument.

Public present or not during the capture

There is no universal rule. On a Gothic cathedral project, the capture was done during a limited morning opening slot, without disturbing the visitors present. At a museum, several scenarios are possible: outside hours (maximum quality), a limited slot (a compromise), or with the public (a testimony of the living place).

Coordination with the site staff

A reference contact on the monument side is useful for opening doors, accessing restricted areas, and safety questions. The capture itself does not require any interruption of activity.


Step 4, processing and quality

Once the capture is in, several hours to several tens of hours of GPU processing. This step is invisible to the client, but determines the final quality.

Production pipeline

Alignment of the images, optimisation of the cloud of Gaussians, visual quality control from several angles, colour adjustments if necessary, integration of derived media (photos, videos, cross-section drawings).

Quality criteria

Three main indicators. Sharpness at close zoom. Navigation consistency (no blind spots or visual “floating”). Chromatic fidelity (the stone, the wood, the stained glass must resemble reality, not a saturated or desaturated version).

Standard delivery time

For a base deliverable, 24 to 48 hours after the end of the capture. For an enriched deliverable (guided path, narration, multilingual), two to four weeks depending on the editorial scope.


Step 5, promotion and distribution

The delivered model must live. Five common uses.

Integration on the institutional site

A dedicated page on the site, in an iframe or via a full-screen link. Linking with existing content (the artwork page, the collection page, the news page, the patronage page).

Guided path with narration

An editorial layer added to the model. Clickable points of interest, audio, historical text, archive photos. This layer turns a model into a mediation tool. It is what we call the Story package.

Derived media

From the 3D model, you extract photos from every angle, path videos, cross-section drawings, high-resolution images for the press or printed materials. This avoids additional photoshoots.

Reference archive for heritage authorities

The source file can be deposited in heritage archiving systems. For listed institutions, it is a piece that adds to the monument’s reference file.

Partner distribution

Tourism office, cultural platforms, media, schools. The web link can be shared with different partners without duplicating cost or technical constraint.


Step 6, maintenance and updates

A monument’s digital twin lives for several years. Three questions to anticipate.

Long-term hosting

Who hosts the model, on what infrastructure, at what recurring cost. Best practices: professional hosting with an availability guarantee, regular backups, transparent terms in the event the relationship ends.

Seasonal or post-restoration updates

A monument evolves. Adding a temporary exhibition, restoring a chapel, changing a visitor path. A new capture can be planned to update all or part of the existing model.

Evolution of the viewers

3DGS web viewers evolve rapidly. A migration to a new viewer can bring useful features (native mobile, AR mode, performance optimisations). Anticipating this evolution from the initial specification makes upgrades easier.


A concrete case, a Gothic cathedral

One of the flagship monuments in our portfolio, a Gothic cathedral (project under a confidentiality agreement), was the subject of a complete capture. Here are the key lessons we take from it, transposable to other monuments.

The light of the stained glass is a character in its own right. It changes minute by minute. The capture slot was chosen to stabilise the light over the duration of the shoot. On a monument with stained glass, never underestimate this parameter.

Verticality demands particular attention. A cathedral is also about height. Scanning the vaults and arches requires an adapted capture method, unlike that of a museum at human height.

Editorial makes the difference. A raw model is pretty. A model with a path, historical points of interest, explanations, archive photos and multilingual narration is a mediation tool. The difference is made after the capture, in the editorial.

Maintenance matters. A monument is viewed over time. The model is now accessible and used. Anticipating hosting, occasional updates, and viewer evolutions is part of a serious project.


How much does a 3D monument digitisation cost?

Three variables shape the budget.

The scope to cover, rooms, grounds, heights, restricted areas.

The level of editorialisation, raw deliverable, guided path, multilingual narration, advanced integrations.

Hosting and maintenance, one-off or long-term, updates included or billed separately.

For an institution with a modest annual budget, entry-level packages exist. For a flagship project with multi-channel rollout, bespoke packages are structured case by case. See our three engagement packages to understand the structure.


To conclude

Digitising a historic monument in photorealistic 3D is no longer an exploratory project. It is a proven, structured and economically viable approach, when conducted step by step with a provider that masters the whole chain, from scouting to long-term hosting.

For heritage institutions in Switzerland considering a project, 2026 is a good time. The technology is mature, the ecosystem stabilised, and the competitive advantage of being among the first to position yourself remains real.

To go further, see our Heritage, culture, tourism sector page, or talk directly about your project via a meeting.

Frequently asked questions

How long does a 3D capture take in a listed monument?

One to three hours for a chapel or a small museum, two to six hours for a medium-sized church, six to ten hours over two days for a cathedral. The capture is non-intrusive and can, in most cases, be done with the public present or outside opening hours depending on the site's constraints.

Do you have to close the site to the public during the capture?

Not necessarily. 3DGS capture is designed to be discreet, with no heavy set-up, no flash, no noise. On busy sites, we often favour limited opening hours or the slot just before or after. On less frequented sites, the capture can take place with the public present without disturbance.

Who owns the delivered 3D model?

The client. The model belongs to you; you can use it on your website, your marketing channels, your mediation materials, your archives. Hosting and maintenance are provided by default, but the source file can be delivered on request to guarantee long-term independence.

Can the 3D model be integrated into our own website?

Yes. The model is viewable via a shareable web link, and can also be embedded in an iframe on your own site (collection page, news page, patronage page). Several viewers are compatible with smooth integration, without degrading your site's performance.

Can the model be archived long term without dependence on the provider?

Yes. 3DGS files are produced in open formats (PLY, SOG) that can be read by several independent viewers. You can request delivery of the source file to archive it in your documentary systems. It is an important guarantee of independence for heritage institutions with long conservation horizons.

What is the delivery time for a heritage digital twin?

24 to 48 hours for a base deliverable after the capture, two to four weeks for an enriched deliverable (guided path, narration, points of interest, multilingual version). The exact time depends on the editorial scope requested and the number of rooms or areas to cover.

Have a project in mind?

Reply within 24 hours.