Meanwhile, Shayne Brandon, the systems administrator and 3-D data acquisition specialist at the Institute for Advanced Technology in the Humanities, was also documenting the building, recording it with photogrammetry, which records objects in single-image frames that are then processed with computer software to produce a 3-D point cloud or model.
Brandon recorded the outside of U-Hall with a quadcopter, a four-rotor, radio-controlled drone carrying a camera. He also photographed the inside of the building with a Nikon D850 hand-held camera. Shortly before Christmas, he took thousands of photos of the ceiling of the arena.
“I am trying to capture the ceiling, the ribs of the dome and the shape of the ceiling and the floor,” he said.
Brandon had already taken 4,000 photos of the exterior of University Hall using the drone and another 17,000 inside. He used a 360-degree camera to record more than 150 spherical panoramas throughout the building and arena.
“I went to any location that I thought might evoke a nice childhood or family memory and some that visitors might have liked to view, but couldn’t,” Brandon said. “I also made changes in lighting and camera height in order to capture the similar views under multiple conditions.”
To create a 3-D model from the images, Brandon uses software running on the Rivanna High Performance Computer Cluster, which determines the surface geometry of U-Hall from the set of photographs, and along the way computes the camera lines and the vantage point from which each was taken. Brandon said the photogrammetry data can be combined with Rourk’s team’s laser scans to create more complete models that can have a variety of uses.
Worthy Martin, acting director of the Institute for Advanced Technology in the Humanities and a member of the computer science faculty in the School of Engineering, said the 3-D data-acquisition methods such as laser scans and photogrammetry record a series of points along a surface and then reproduce the shape in a point cloud, similar to a pointillist painting.
“Will is an artist in knowing where to put the laser scanner and how to put them together, the same way Shayne is an artist in regard to the photogrammetry, which is a different process,” Martin said. “It shows real artistry in a space as complex as those hallways are, because you have to set the scanner so there is a direct line of sight from the tripod to every point you are going to have a data point for. Picking a place from which you actually see all the surfaces is very much an artist’s task.”
Martin said while the interior could be scanned with a laser, the exterior roof could not, since the lasers are ground-based and not able to “see” the curved roof. He said the drone images will be used to reproduce the roof.
“To get a model for the exterior, we are going to have to combine the data sources from several things,” Martin said, “There is an artistry in taking data from the lasers and data from photogrammetry and putting those together and trying to make a coherent, even point cloud.”
The data from Brandon’s photogrammetry will be processed together with Rourk’s laser scan data to produce a combined dataset.