Android book-scan app tames untidy tomes
Dewey Decimal System sorts your shelves
An augmented-reality application can look at a shelf of books and tell you which ones are out of order, for the professional librarian or obsessive compulsive bibliophile.
The books have to be tagged first with a machine-readable label, but once that's done the user can walk down the shelves looking for red crosses in a sea of green ticks, which change to a green tick showing which direction the problem book should be moved in to put the collection back in order.
Attaching all those labels is the obvious problem. The app can't read the often poorly-attached numbers - Dewey Decimal references - that adorn the spines of library books. Instead it requires an entirely new tagging system that can be read with a phone's camera and processed by an Android application.
That might be too much for a library, which will have to judge the cost of labelling every book against the expense incurred by having one in the wrong place; but it won't put off obsessive bibliophiles who know you can't put a price on perfection.
The Chronicle of Higher Education has more details of the application, and some thoughts about how to make it commercially viable, but the following video will probably tell you all you need to know:
If your first-editon Gutenberg Bible is sitting on a shelf amongst a plethora of other books you have a bigger problem than a possible sticky label :p
RFID not nec that fab for shelf order
RFID, when used in Libraries is great, however it seems to be built up as some sort of magic bullet for all your Libraries issues where in reality it is good for some things (self check etc) and very ordinary for others (security for instance).
On to the subject at hand, while it can be done, most of the RFID scanners purchased by Libraries are bought for Inventory and stock control (or looking for specific items on a shelf) and are not necessarily that accurate for checking shelf order. If you want to process a whole shelf load of items quickly, using a portable RFID scanner is great, likewise finding an individual item on a shelf and its general location quickly is where they can shine too. Checking shelf position (especially of thin items) is not nearly so good.
ISBN is not enough
Nobody, but *nobody* shelves their library in ISBN order. By author, by title or by subject would be good. The British Library makes the British National Bibliographic datasets available for free (registration required; non-commercial use), see http://www.bl.uk/bibliographic/datafree.html.
Note that the Dewey Decimal Classification is a proprietary system which you must licence from OCLC.
... to the cover of my first-edition Gutenberg Bible...?
RFID microtags FTW.
I've been wanting to do this for ages.
Look for straight lines to identify books whether vertical or horizontal.
Take biometric type info, eg: identifiable text (ocr'd if possible), identifiable barcodes (scanned if possible), identifiable font types, identifiable coloured regions, distances between identified items, etc.
Log "fingerprint" against book.
Build robot to find book.