Book-Scanning Robots Digitize Delicate Texts

21st-century robots read 16th-century Bavarian books

PHOTO: TREVENTUS MECHATRONICS

A DELICATE TASK

A pair of extra-gentle digitizing robots scan books at the Bavarian State Library.

Students come from the world over to study the Bavarian State Library’s collection of works from the time of Martin Luther. But this year, the Munich institution’s 450th in existence, the most voracious readers of its ancient collection will be a pair of robots. In the library’s basement, two machines called ScanRobots are whirring away at 700 pages per hour and are scheduled to digitize all of the four­century-old books in the library, some 7.5 million pages’ worth, by 2009; the scanned books will be put online.

”This is the entire knowledge of this period”
—Markus Brantl, director of the Bavarian State Library’s digitization center

Markus Brantl, director of the library’s digitization center, says it’s vital to digitize unique content like the 16th-century collection, both for preservation purposes and to open access for readers, academics, and laypersons alike. ”This is the entire knowledge of this period, [from] theology to mathematics—everything,” he says. But in making the material much more widely available, the Bavarian library is also giving a boost to a scrappy band of robot engineers from Austria’s Vienna University of Technology, who are out to upset the scanning market.

The ScanRobots are the debut project for Treventus Mechatronics, which spun out of the university after Professor Wolfgang Zagler sketched out the idea for the machines on a train ride in 2002. What makes the robots unique is their ability to scan books that are opened only to a 60-degree angle, which keeps the spines and pages of older books from being damaged by the strain of being fully opened. The US $125 000 machines work by holding the pages in place with soft suction and moving a scanning head vertically while the book is held open underneath. The scan head contains two 30-degree glass prisms that project onto high-quality cameras. The head doesn’t touch the page—the scanning camera records both open pages through the prisms simultaneously in high resolution as it methodically moves up and down, like an oversize sewing machine needle. Gentle air jets turn each page after the scan is complete. Nightmares of sucking pages out of the Gutenberg Bible will remain just that, idle anxiety dreams, promises Stephan Tratter, a former grad student at the university and now head of marketing and R&D at Treventus.

Tratter says the firm got off the ground only after winning a few European research contests for seed money. Its first robot proto­type went through trials at the University of Innsbruck, in Austria, last spring, and the firm now has robots at three other sites.

Even though the ScanRobot doesn’t have legs or wheels, Tratter says it is indeed a robot and not just a glorified photocopy machine. ”This one turns the pages automatically. In principle, it has no need of an operator,” he says, although with such delicate projects as the one in Munich, there will be one. ”The robot is also able to recognize errors and learn not to make them again.”

Brantl is glad to hear that. His digitization department has been working for 10 years and now has 23 000 books online. The ScanRobots will surpass that figure this year alone. The library ordered the Austrian book-bots last fall to handle delicate digitization work like the 16th-century project. The Bavarian State Library, one of the ­largest in the German-speaking world, has already cast its lot with Google in the sometimes controversial struggles over digitizing and ­posting online the vast volume of literature that no longer is covered by copyright law. When the job is done, says Brantl, the Bavarian library will have digital versions of 1 ­million of its 9 ­million ­possessions, including its 80 000 ­medieval manuscripts and 20 000 incunabula (printed books made before 1501).

Advertisement
Advertisement