When the University of San Francisco set out to reassemble a historic Portuguese portal from a pile of stone fragments, they knew they’d found a hard problem. Surprisingly, while the fragments they received were carefully labeled, there was no key for their reassembly. Worse, several sets of numbers were evidently being used as labels. Traces of different markings could be seen, often conflicting with others. Without a definitive guide to the assembly, the University was left to fit the pieces together themselves. Unfortunately, placement for many of the small stone pieces were ambiguous to laypersons and experts alike. Even when it was clear that a piece belonged to a recognizable feature (for instance, the center keystone arch), the neighbor pieces would stubbornly refuse to reveal their proper locations.
Working with large stone blocks like puzzle pieces is not a new challenge, and it has never been easy. Strolling at night through the brightly lit Temple of Luxor in Upper Egypt, it’s possible to spot places where mid-20th Century masons unintentionally inverted a piece before cementing it in place, forever. The goal of the USF project was to avoid those kinds of errors by testing conjectural assemblies on the computer before working with the physical stones.
Fitting the puzzle pieces
In order to build a digital model of the portal, first USF needed to capture each stone piece using laser scanning. CyArk Project Manager Justin Barton supervised the digital capture of the fragments, which were laid out in crowded rows in a campus storage facility. The scanning itself took only a few days — an impressive feat given the requirement that each of the 50+ stone chunks needed to be sampled from all sides. Once the raw data was processed, the INSIGHT team began to register the data, creating a separate polygon mesh for each of the physical stone pieces.
Next, we at INSIGHT began to put the pieces together. Fortunately, many of the largest pieces were easy to orient — they could act as anchors around which the other pieces could be strung. Jesuit Father Thomas Lucas, USF professor of art and architecture, supervised the fragment reassembly directly in the INSIGHT office. Jun Nagaoka from the INSIGHT team worked with Father Lucas in iterations, suggesting joins and testing connections. Once the main pieces were placed, revisions were hashed out in daily images attached to e-mails. Two weeks were spent in this puzzle piecing mode. Nearly all of the time was not put into the technical requirements of digital modeling, but instead was spent thinking.
With the approved digital template in hand, Steve Bowers and the team at Chapiter Construction Company carried out the physical construction, from orthographic drawings made by INSIGHT (below, left). As shown above, ID information was transferred from the digital model to the physical fragments being assembled. One of the unexpected advantages of the digital model for Chapiter was the ability to accurately take into account the uncut stones protruding on the back side of the assembled portal. As seen in the top, right image, the back side of the portal is a complex assembly that required careful consideration for anchoring and attachment.
This was not the first stone portal INSIGHT has been asked to help re-erect from fragments. Long-time INSIGHT board member Tom Lewis completed a similar stone portal reassembly at Santana Row, San Jose, in 2006. With Jamie Hyneman and Kevin Cain, Tom also also reassembled part of a portal and romeo balcony for a Berkeley residence designed by Jo Mora in 2004.