A Stephen F. Austin State University team is creating a printable replica of an Apollo 14 Moon rock for travelling education by the Naranjo Museum of Natural History. The NASA sample must remain inside a reflective acrylic pyramid, so geometry has to be captured without touching it.

Why straightforward scanning failed

Two high-performance scanners could not produce reliable geometry because the acrylic reflects and refracts measurement light while obscuring surface detail. The team shifted to controlled, multi-view photography and photogrammetry, and is comparing AI tools that reconstruct 3D models from images.

Photogrammetry estimates cameras and a point cloud from matching surface features. In a sealed case, reflections move differently from the object and can create false points or floating surfaces. The solution is not one magic algorithm but careful lighting, masking, image selection and manual review.

From raw mesh to printable object

Reconstruction leaves holes, self-intersections, noise and uncertain dimensions. Art, computing and library specialists are collaborating on the mesh. It must become watertight, sufficiently thick and orientable for printing without inventing uncertain detail with false precision.

Replica fidelity has layers. External shape and scale can be measured; colour can be approximated from photographs. Composition, mass, microscopic texture and scientific value cannot transfer to a printed object. Labelling must identify the result as a replica, never the original sample.

Appropriate verification

The team can use scale references, independent image sets and projection of model silhouettes onto photographs. Rescanning the printed object and comparing it with the digital mesh measures manufacturing accuracy, but uncertainty relative to the enclosed original must still be reported separately.

Museum value and limits

A handling replica makes geometry tangible, reduces transport and insurance risk, and can improve access for visitors with visual impairments. A versioned digital file can be reproduced at education venues, provided provenance is retained so later corrections do not become confused with earlier models.

Editorial conclusion

The project's main lesson is not AI alone but a multi-method workflow that documents uncertainty. A printed replica is a strong museum tool when it is durable enough to handle and honest enough not to claim more about the original than the photographs support.

Sources

- Stephen F. Austin State University / 3D Printing Industry: SFA team uses photogrammetry and AI to replicate a Moon rock — institutional project account and specialist report, 21 September 2026. - NASA: Apollo 14 mission and lunar sample background — official mission context.