3-D Photogrammetry

Photogrammetry has been featured in ArtLens Gallery since 2019. In the ArtLens Exhibition, projected interactive games use gesture sensing that responds seamlessly to body movement and facial recognition for an immersive experience. In each gesture-based game, all projected 3-D artworks can be rotated and viewed at all angles. These 3-D models were created using a technique called photogrammetry.

a view of a vase in CMA's collection

Photogrammetry is a way to visually represent real-world objects in a manner that combines photographic imaging with the metrics of scale. The CMA’s digital innovation team led the implementation of photogrammetry to create 3-D models of objects. This is not a simple task; it represents true innovation in the museum field. This installation features numerous objects that were challenging to capture using photogrammetry due to their reflective materiality and intricate details. However, we embraced the opportunity, and the museum is now recognized and respected in the photogrammetry community.

The method combines photographic imaging with the metrics of scale to show both the shape and the texture of an object with a high level of photorealism. The process starts in the CMA’s photo studio, where a photographer takes a series of overlapping photographs capturing every part of the object several times. Those images are then fed into a 3-D software program, which, after a series of steps, produces the 3-D model. This is then taken through postprocessing, conducted by a photogrammetry specialist, to clean up and perfect the details to make the model replicate the object with a high level of accuracy.

The photogrammetry method we used was developed by senior photogrammetry experts at the US Bureau of Land Management. It optimizes the data sets utilized and incorporates individually calibrated scale bars that are accurate to 1/10th mm or better. So the resultant 3-D models produced with this method are accurate to that level and have the visual appearance of the actual objects. This method contrasts with laser scanning, which can produce accurate 3-D models, because the resulting objects have a blank surface appearance.

Objects such as the fifth-century Greek black figure amphora were photographed using polarized light to eliminate reflections (which confuse the software) and against a black background to further dampen reflections from the surrounding set.

a graphic showing how different camera angles are used to capture images of the object
A graphic representation shows camera positions. Blue squares are all the camera angles captured.

To create a 3-D model using photogrammetry, a series of overlapping photographs are taken of an object, showing every part of the object several times. Those images are then fed into a 3-D software program, which, after a series of processing steps, produces the 3-D model. To complete this work, we used Agisoft Metashape Pro, formerly Agisoft PhotoScan Pro. While some objects were photographed in the galleries or conservation labs, most were brought into the CMA’s photo studio and photographed on turntable setups that were custom configured for each object to optimize results. We used a Crayfish 60 turntable from Seabass 360, a small firm in London, England. This well-built turntable has an excellent control interface supporting up to 286 pounds. It also features an overhead rotator configuration for hanging certain objects, allowing them to rotate for the camera. For larger objects, we used a heavier manual turntable built by the museum’s exhibition production staff.

three side by side images showing the process of photogrammetry
Image 1. Solid in processing state. Image 2. Sparse cloud in processing state. Image 3. Textured, full-photographic surface information in processing state.

With a track record of more than a decade of digital transformation, the CMA has invested in iterative, mindful digital initiatives that expand access and connect people to the collection. A major facet of their success is the digitization of the collection, which stands at 98% and is critical to major projects like the comprehensive Open Access initiative that allows the public to share, collaborate, remix, and reuse high-resolution images of more than half of the collection. In addition, the digitized collection has been the launchpad for audience-centered web-based collection tool sets such as ArtLens AI and ArtLens for Slack and for interactives in the ArtLens Gallery. 

Photogrammetry augments the projects enabled by digitization. Three-dimensional models add value to research conducted by conservators and curators, preserve objects, and make it easy to view and handle objects that are large, delicate, or complex. As part of the museum’s newest strategic plan, the CMA is expanding its holdings of digital assets like these, which are vital for future initiatives. The museum is focused on using new technology to create innovative experiences that help bring art to life, including but not limited to forays into the metaverse for education and entertainment, increased access to resources on repositories across the web, and future immersive exhibitions. We’ve seen that photogrammetry facilitates engagement with the collection in ArtLens Gallery, on Collection Online, and in immersive exhibitions like Revealing Krishna. In the next five years, we hope to have one thousand 3-D assets in the collection to expand and iterate on these initiatives.

Learn More about Photogrammetry