Why it matters
Without such mapping, interventions risk irreversible damage, misinterpretation of surface changes, missed losses, and disputes over insurance or loan readiness due to opaque condition records.
From design to safety
Design
AI guides treatment planning
AI reviews baseline condition data, prior studies, proposed methods, pigment maps, varnish layers, binder types, and craquelure morphology to generate a risk map, imaging plan, and material choices for treatment proposals; it suggests reversibility checks and conservation ethics notes.
- IR camera
- RTI dome
- UV lamp
- Digital stereomicroscope
Build
Workflow and equipment integration
Data pipelines establish metadata schemas (Dublin Core, PREMIS), calibration routines for color and geometry, and protocols for capturing craquelure, cupping, flaking, and losses; validated outputs ensure reproducible reports across studios and loan projects.
- IR camera
- RTI dome
- UV lamp
- Digital stereomicroscope
Operation
Live mapping during examination
AI fuses imaging streams (raking light, UVF, XRF start-map) to produce dynamic damage maps, craquelure density, and loss probability; updates treatment proposals, documentation, and condition reports with concurrent QA checks and provenance stamps.
- IR camera
- RTI dome
- UV lamp
- Digital stereomicroscope
Maintenance
Model and instrument calibration
Regular checks verify alignment between imaging data, color calibration, and AI defect scoring thresholds; versioned outputs, calibration certificates, and archival of instrument drift data ensure long-term traceability.
- IR camera
- RTI dome
- UV lamp
- Digital stereomicroscope
Safety
Safety and ethics governance
COSHH assessments, object safety protocols, and PACR accreditation controls ensure safe use of UV exposure, solvents, and data handling; sign-off by accredited conservator and studio management with training records.
- UV lamp
- RTI dome
- XRF spectrometer
- Digital stereomicroscope
Part by part
- 1
IR camera
- Watch for
- Low signal in shaded craquelure regions or reflective varnish masking cracks.
- AI action
- Detect crack networks and flag inconsistent illumination for re-imaging.
- 2
RTI dome
- Watch for
- Inadequate multi-angle surface texture capture leading to missing microcracks.
- AI action
- Generate high-resolution 3D texture map and export to treatment planning report.
- 3
UV lamp
- Watch for
- Fluorescence variability across varnish layers indicating varnish age or composition changes.
- AI action
- Trigger targeted imaging with updated wavelength settings and save spectral map.
- 4
Digital stereomicroscope
- Watch for
- Small-lost-to-view paint flakes or underlayers not visible at macro scale.
- AI action
- Capture zoomed stacks and produce microcrack density metrics.
- 5
XRF spectrometer
- Watch for
- Uncharacteristic elemental signatures near losses suggesting overpaint or past restorations.
- AI action
- Annotate material map and flag areas for confirmatory microscopy or sampling notes.
- 6
Portable paint loss mapping tablet
- Watch for
- Discrepancies between map overlay and image files across sessions.
- AI action
- Synchronize versions, log changes, and generate a consolidated loss map.
Worked example
Catching a craquelure fault before a charter
Illustrative scenario with sample data.
An easel painting with age craquelure, cupping, and scattered paint losses is prepared for loan; AI maps surface condition to inform treatment proposals.
Signal in
pre-treatment condition report; high-res images: craquelure_map.png, cupping_map.png; XRF readings: Pb, Zn; varnish thickness: 12 μm; pigment maps: Prussian blue, Naples yellow; binder ratio: 65% linseed oil, 35% stand oil
System out
Damage map with density scores, probabilistic loss zones, recommended imaging sequence, treatment proposal draft, and necessary safety notes.
Human sign-off: Head of Conservation reviews the damage map, approves imaging plan, and signs off on the treatment proposal.
How to roll it out
Phase 1
Weeks 1 to 4
Baseline scanning, calibrations, and initial AI model validation with existing condition records; implement metadata schema and instrument checklists.
Phase 2
Weeks 5 to 8
Pilot runs on active projects, refining thresholds, reporting templates, and audit trails within the studio workflow; validate reversibility checks.
Phase 3
Weeks 9 to 12
Full deployment across projects, formalizing archival procedures, version control, and staff training in imaging and ethics.
Where people stay in charge
- AI-generated decisions must be reviewed by an accredited conservator before treatment proposals are finalized.
- All imaging and material data must be stored with verifiable provenance and versioning signed off by studio management.
- Critical safety decisions, such as UV exposure limits and solvent use, remain under qualified personnel control.
Questions owners and crews ask
What is craquelure mapping in painting conservation and how does AI help?
Craquelure mapping identifies crack patterns and density, while AI integrates imaging data to produce a consistent map that informs treatment decisions and documentation.
How does AI handle paint loss mapping on easel paintings?
AI aligns photography, raking light, and spectral data to locate losses, quantify their extent, and help plan fills and inpainting within reversible methods.
Who approves AI-generated treatment plans?
An accredited conservator reviews and signs off the treatment proposal, condition report, and archival records tied to the AI outputs.
What records are created for insurance and loans?
Damage maps, imaging logs, material maps, and treatment proposals are compiled into condition reports and archives to support loan and insurance processes.