7AAVDH22 Chap.3 Capture Planning and Paradata
Capture Planning and Paradata
Translate the heritage claim into a capture goal
Photogrammetry begins before the camera moves. Identify the feature, scale or context that the model must support. Manufacturing marks may require close, stable views; spatial relation may require environmental context; handling traces may be weakened by overly smooth lighting or cleanup.
The capture goal determines which surfaces matter and which uncertainty would undermine the intended use.
Plan connected coverage
Adjacent photographs need shared features while viewpoints change enough to describe form. Organise overlapping passes around the object, include elevated and lower views where safe, and mark surfaces that remain hidden.
More images do not guarantee better evidence when they repeat the same weak angle. A coverage audit should identify connected regions, isolated groups and absent surfaces before processing begins.
Control light, focus and scale
Stable, diffuse lighting can reduce moving highlights, while reflective, transparent, dark or featureless surfaces may still resist matching.
Record exposure and focus decisions that affect comparability. Include scale and orientation information when size or spatial relationship matters. Site constraints, handling rules and public activity may limit the ideal setup; those constraints must be recorded because they shape the model’s evidential boundary.
Separate optical and geometric failures
Glare, blur and exposure changes affect feature recognition.
Missing viewpoints and movement produce different geometric gaps. Diagnose the cause before recapturing. Repeating the same path under unchanged conditions can increase file count without improving evidence. A targeted repair changes the earliest failed condition while respecting object and location safety.
Write useful paradata
Metadata identifies files, dates, formats and devices.
Paradata explains consequential decisions and transformations. Record when the object was moved, lighting changed, images excluded, geometry filled or surfaces cleaned. Explain why the decision was made and what claim it affects. The goal is not a diary of every command but a trace that lets another person judge fitness for purpose.
Prepare a capture handover
Keep untouched originals and traceable derivatives.
Group files by orientation or setup, identify scale and reference views, mark known gaps, attach permissions and list checks for reconstruction. A strong handover states what the data set can support and what it cannot. This prevents later modelling from hiding missing evidence behind a clean surface.
What this chapter covers
- 01
Capture Goal
- 02
Image Overlap
- 03
Surface Constraint
- 04
Scale Reference
- 05
Paradata
- 06
Coverage Audit
Worked example · free
Repair capture of a reflective figure
- 2Restate the heritage feature and failure causes.
- 3Plan lighting, overlapping views and scale.
- 3Record remaining gaps and restrict the claim.
Key terms
- Image Overlap
- Image Overlap is shared visual evidence between changing viewpoints that helps connect photographs during reconstruction.
- Coverage Audit
- Coverage Audit checks connected views, isolated groups, hidden regions and whether important surfaces have sufficient evidence.
- Scale Reference
- Scale Reference records a known dimension or calibrated relation so the reconstructed model can retain meaningful size.
- Difficult Surface
- Difficult Surface is reflective, transparent, dark, featureless or moving material that weakens stable correspondence across photographs.
- Paradata
- Paradata explains significant capture and processing decisions, their reasons and their consequences for interpretation.
Capture Planning and Paradata FAQ
How many photographs are enough for capture?
There is no useful universal count. The image set needs connected overlap, adequate viewpoints for important surfaces, stable appearance and the scale required by the intended use. Audit coverage and uncertainty rather than treating file quantity as evidence of quality.
What should be recorded as paradata?
Record decisions that change evidence or interpretation: object movement, lighting changes, excluded images, failed alignment, generated geometry, texture cleanup, scale decisions and known gaps. Link the entry to a stage or file and explain the consequence.
What if an object cannot be fully photographed?
Respect handling and site constraints, capture what is safe, and document inaccessible regions. Narrow the model’s intended use and avoid filling unseen geometry as if it were observed. A visible, explained gap is more defensible than an undocumented plausible surface.
Assessment move
Sketch a camera orbit around an unfamiliar object and label overlap, elevated views, scale, reflective areas and hidden regions. Then write three paradata entries for a lighting change, an excluded photograph and an inaccessible surface. For retrieval, explain the difference between metadata and paradata without examples, then apply the distinction to your own capture package.
Finish with a fitness-for-purpose sentence that names both an adequate use and a prohibited inference. Compare two capture plans using the same photograph count but different viewpoint distributions, and explain which better supports connected evidence. Rehearse the handover aloud by naming originals, derivatives, permissions, uncertainties and the first reconstruction check.
Add a no-recapture scenario so the response distinguishes repairable processing choices from permanent gaps in the record. Verify that every exclusion is recorded before opening the reconstruction software.