Two years of planning (1997-98)

Photographs and text by Marc Levoy
May 24, 1999, and October, 1999

The Digital Michelangelo Project began in January of 1997. The 20 months from then until we shipped our equipment to Florence in September of 1998 were spent ramping up our staff, developing hardware and software, and generally laying the logistical groundwork required to mount a large-scale 3D digitization effort in a foreign country. Here are some photographs from this "pre-production" phase of the project.

Our first few months were spent designing scanners on the backs of napkins and traveling back and forth between the United States and Italy. During these trips, we presented our project to the Italian authorities, obtained permissions to scan the statues, and gathered measurements that we would need when building the scanners. One crucial measurement - the approximate height of the David - was taken from art history books instead of from field measurements, and it proved to be wrong. (The David is nearly 3 feet taller than the books say he is!) In this photo, taken in July of 1997, Marc Levoy and Brian Curless measure the scattering of laser light from the leg of the David. We suspected that controlling laser scatter would be important when scanning marble statues. (Later events proved us right.)

The main hardware component of our project was a laser triangulation scanner and motorized gantry customized for scanning large statues. We developed this device in two steps, building a prototype scanner head first in order to test out our ideas about spatial resolution, working volume, standoff distance, and scanning pattern. Here is the prototype, manufactured to our specifications by Cyberware. The object being scanned is a full-size replica of an Egyptian sarcophogus from the British Museum. (This replica was made by Industrial Light and Magic for Raiders of the Lost Ark.) The laser light and camera line of sight were added in Photoshop to enhance the visualization. The scanner head is mounted on a non-motorized photographic camera stand for testing.

Here is an early design model for the motorized gantry on which the scanner head was to ride. This model was built by Marc Levoy and Benjamin Levoy (age 6) from K'nex parts. It consists of a vertical truss, a horizontal arm that translates vertically on the truss, a pan-tilt head that translates horizontally on the arm, and a scanner head. It is an accurate scale model (2.5:1), and the scanner head, shown sitting atop the horizontal arm, is motorized (2 size-A batteries). A laser pointer (not shown) is mounted to the scanner to simulate the scanning pattern. Click here to see photographs of our final scanner and gantry.

In addition to our Cyberware laser triangulation scanner, we also planned to take with us to Italy a jointed digitizing arm and small triangulation laser scanner made by Faro Technologies and 3D Scanners Ltd. and this beta model of the Cyrax time-of-flight laser range scanner, manufactured by Cyra Technologies. It can scan objects up to 100m away with an accuracy of about 5mm. Our plan was to use it to scan architectural settings. Our version was to be outfitted with a high-resolution digital color camera (not shown here).

At the same time we were designing and building hardware, we were also trying to decide what to scan. During his lifetime Michelangelo created nearly 50 statuary works, not all of which are even in Italy. From the outset of the project we realized that creating a complete catalogue in a year's time abroad would be impossible. We therefore looked for statues and spaces, preferrably in Florence or Rome, that posed unique technical challenges or provided unique historical opportunities. One space we particularly wanted to scan was the interior of the Medici Chapel in San Lorenzo in Florence. The architecture is by Michelangelo, and seven of the nine statues are believed to be by his hand. Although the work is incomplete, it is neverthless considered to be the crowning achievement of his middle years. Using the Cyberware and Cyrax scanners, our plan was to create a 3D computer model of the entire room, a high-resolution model of as many of its statues as time permitted, and a light field of at least one statue group. Click here to read about our scan of the chapel.

Although most of our plans worked out, there were several disappointments. One of these was the Pieta in St. Peter's Basilica. The logistical difficulty of scanning this work was formidable: it is mounted high on a pedestal, it is sandwiched between an immovable altar in front and a niche behind, and it is polished to a high sheen. As a result, we spent a good deal of time looking for a cast to scan. Here is a photographic comparison we made of the original and a cast we located at a carving studio in the town of Pietrasanta. The cast is good, but it's not good enough at the scale we'll be scanning (0.25mm). Look for example at the stigmata and at the spaces between the fingers. Of course, these features would also be difficult to scan. In the end, we decided to scan the original. Unfortunately, although the Vatican had already permitted us access to the original (for example to take this closeup photograph), they could not guarantee us enough hours alone in the Basilica to scan the entire statue. Click here to see images of everything we actually managed to scan.

(Photograph courtesy of Prof. Rodriguez-Almeida.) While in Italy, we also planned to tackle several side projects. One of them was to scan the 1,163 fragments of the Forma Urbis Romae, the giant marble map of ancient Rome. Here is a photograph of four typical fragments. They are about 3 inches thick and together weigh about 50 pounds. These fragaments were fit together in the early 1980's by Emilio Rodriguez-Almeida, the greatest living expert on the map. As this example shows, the matches between fragments are not obvious from an examination of their top surfaces, which are often eroded. These particular pieces do fit, however, as can be verified by examining the edges, largely hidden in this photograph, where they touch each other. It is these edges that we planned to scan. Click here to read about our scan of the map.

Heroes of the planning process

Getting permission to scan Michelangelo's statues was a long, delicate, and occasionally painful process. The story has heroes, villians, and moments of high drama. We won't tell the whole story here, but we'd like to identify two of the heroes: Christina Acidini Luchinat (at left), 2nd in command at the Superintendency of Fine Arts in Florence, and Franca Falletti (at right), director of the Galleria dell'Accademia, home of Michelangelo's David. These two women, recognizing the potential impact of our project, took it under their wing and made it their own. They helped us shape its goals, they shepherded it through the Italian bureaucracy, fighting opposition when it arose, and they built for us an amazing team of local supporters and collaborators. We can't thank them enough.

webmaster@graphics.stanford.firenze.it

	Our first few months were spent designing scanners on the backs of napkins and traveling back and forth between the United States and Italy. During these trips, we presented our project to the Italian authorities, obtained permissions to scan the statues, and gathered measurements that we would need when building the scanners. One crucial measurement - the approximate height of the David - was taken from art history books instead of from field measurements, and it proved to be wrong. (The David is nearly 3 feet taller than the books say he is!) In this photo, taken in July of 1997, Marc Levoy and Brian Curless measure the scattering of laser light from the leg of the David. We suspected that controlling laser scatter would be important when scanning marble statues. (Later events proved us right.)
	The main hardware component of our project was a laser triangulation scanner and motorized gantry customized for scanning large statues. We developed this device in two steps, building a prototype scanner head first in order to test out our ideas about spatial resolution, working volume, standoff distance, and scanning pattern. Here is the prototype, manufactured to our specifications by Cyberware. The object being scanned is a full-size replica of an Egyptian sarcophogus from the British Museum. (This replica was made by Industrial Light and Magic for Raiders of the Lost Ark.) The laser light and camera line of sight were added in Photoshop to enhance the visualization. The scanner head is mounted on a non-motorized photographic camera stand for testing.
	Here is an early design model for the motorized gantry on which the scanner head was to ride. This model was built by Marc Levoy and Benjamin Levoy (age 6) from K'nex parts. It consists of a vertical truss, a horizontal arm that translates vertically on the truss, a pan-tilt head that translates horizontally on the arm, and a scanner head. It is an accurate scale model (2.5:1), and the scanner head, shown sitting atop the horizontal arm, is motorized (2 size-A batteries). A laser pointer (not shown) is mounted to the scanner to simulate the scanning pattern. Click here to see photographs of our final scanner and gantry.
	In addition to our Cyberware laser triangulation scanner, we also planned to take with us to Italy a jointed digitizing arm and small triangulation laser scanner made by Faro Technologies and 3D Scanners Ltd. and this beta model of the Cyrax time-of-flight laser range scanner, manufactured by Cyra Technologies. It can scan objects up to 100m away with an accuracy of about 5mm. Our plan was to use it to scan architectural settings. Our version was to be outfitted with a high-resolution digital color camera (not shown here).
	At the same time we were designing and building hardware, we were also trying to decide what to scan. During his lifetime Michelangelo created nearly 50 statuary works, not all of which are even in Italy. From the outset of the project we realized that creating a complete catalogue in a year's time abroad would be impossible. We therefore looked for statues and spaces, preferrably in Florence or Rome, that posed unique technical challenges or provided unique historical opportunities. One space we particularly wanted to scan was the interior of the Medici Chapel in San Lorenzo in Florence. The architecture is by Michelangelo, and seven of the nine statues are believed to be by his hand. Although the work is incomplete, it is neverthless considered to be the crowning achievement of his middle years. Using the Cyberware and Cyrax scanners, our plan was to create a 3D computer model of the entire room, a high-resolution model of as many of its statues as time permitted, and a light field of at least one statue group. Click here to read about our scan of the chapel.
	Although most of our plans worked out, there were several disappointments. One of these was the Pieta in St. Peter's Basilica. The logistical difficulty of scanning this work was formidable: it is mounted high on a pedestal, it is sandwiched between an immovable altar in front and a niche behind, and it is polished to a high sheen. As a result, we spent a good deal of time looking for a cast to scan. Here is a photographic comparison we made of the original and a cast we located at a carving studio in the town of Pietrasanta. The cast is good, but it's not good enough at the scale we'll be scanning (0.25mm). Look for example at the stigmata and at the spaces between the fingers. Of course, these features would also be difficult to scan. In the end, we decided to scan the original. Unfortunately, although the Vatican had already permitted us access to the original (for example to take this closeup photograph), they could not guarantee us enough hours alone in the Basilica to scan the entire statue. Click here to see images of everything we actually managed to scan.
(Photograph courtesy of Prof. Rodriguez-Almeida.)	While in Italy, we also planned to tackle several side projects. One of them was to scan the 1,163 fragments of the Forma Urbis Romae, the giant marble map of ancient Rome. Here is a photograph of four typical fragments. They are about 3 inches thick and together weigh about 50 pounds. These fragaments were fit together in the early 1980's by Emilio Rodriguez-Almeida, the greatest living expert on the map. As this example shows, the matches between fragments are not obvious from an examination of their top surfaces, which are often eroded. These particular pieces do fit, however, as can be verified by examining the edges, largely hidden in this photograph, where they touch each other. It is these edges that we planned to scan. Click here to read about our scan of the map.