For all the discussion and debate about Google's controversial (and potentially anti-competitive) plan to digitise millions of books, little has been known about how the company is actually doing it.
The audacious plan involves scanning many millions of books, through a system that Google has said is light years ahead of the flatbed scanners we're all used to - systems that inevitably proved unreliable or destroyed the books they were intended to preserve. But Google's never been keen to share the details of how it manages this feat: even when I visited Oxford's Bodleian Library a couple of years ago to find out more, I wasn't allowed to see the scanning room because the machines were considered trade secrets.
Now, though, NPR has managed to uncover the details behind the company's patented book scanning system - and found that the secret is 3D infra-red cameras:
Google created some seriously nifty infrared camera technology that detects the three-dimensional shape and angle of book pages when the book is placed in the scanner. This information is transmitted to the OCR software, which adjusts for the distortions and allows the OCR software to read text more accurately. No more broken bindings, no more inefficient glass plates.
Cunning.
