I remember reading about Google's book scanner that operated automatically using vacuum pressure to gently flip pages. I'd love to see an open source variety of that.
And you can order a bound paperback book made from the scans of just about any out-of-copyright google book. From the Harvard Bookstore (or anyone else that has an Espresso Book Machine):
I don't know where to find a picture of it, but I remember seeing a video of one of their book scanners. Or maybe it was just one version of them.
Picture a wedge (shaped like the V of a book lying open on a table) that could move up and down. The wedge would descend and insert itself into the V of the book's pages, then rising up would suck the 2 left and right sheets against the sides of the wedge, scanning as it went.
Then it would blow those 2 pages to the left (or right I forget) and descend again and do the next set.
I thought it was pretty cool. Have never seen something like it since.
Automated page turning is an incredibly complex problem to tackle. It won't gain you as much as you'd think either. "Book" and "page" are a surprisingly difficult to define categories.
If my experience helping to build the diybookscanner.org project taught me anything, it's that picking the low hanging fruit of small efficiency improvements to the (semi-)manual process is so much more effective...
Which is probably why some of the scanning locations used something very similar to this DIY effort (I'm familiar with the Oxford, UK location that existed in the mid 00s). Humans turned the pages with the finger tips mentioned in another comment.
