High Speed Book Scanner From Trash

book_scanner

[Daniel] sent us his entry to the Epilog laser cutter challenge on instructables. He made a book scanner, mainly out of found parts. The bulk of the project was salvaged from dumpsters, though if you’re not comfortable with that, the free section of craigslist might be able to do the job. The cameras are loaded with CHDK, using StereoData maker, and custom software to compile the images into PDFs. They did a fantastic job of documenting every step of the construction, including helpful tips for some of the more complicated parts. There are several videos in the instructable, so be sure to check them out. We’re particularly amused by the extra step of making the photo captions visually interesting. At 79 steps, it’s a long read, but well worth it.

28 thoughts on “High Speed Book Scanner From Trash

  1. Thanks for posting this!

    I would love a page-turner, too, but that would probably be a whole new Instructable and a hell of a lot of work. Scanning a 400 page book takes about 20 minutes in my device, and using a standard flatbed scanner, it takes 4-5 hours. I think that’s a pretty substantial savings for so little effort.

    I plan to digitize most of the old Navy books that teach tools, electronics, and machining so we can have an open library of basic technique. With a couple of us working, we could have it done in no time.

    If anyone plans to build one of these, please contact me and I’d be happy to help.

  2. This is awesome. I admit I haven’t read it yet, as I’m in the middle of moving home to the US, but I have been looking forever for an easy way to do this, instead of pay someone else and have them remove the pages from the book.

    I have a lot of large, many paged books on horology, and I haven’t read them all yet- to be able to scan them and read them on my laptop when I get time would greatly help me, especially since I’m continually jumping around different countries. I’ve paid several hundred dollars to ship a small part of my library home from Japan, and something like this, in the future, would really help me read without bringing my several hundred kilos of books with me!

  3. Thanks, once I get back to the usa, assuming I can afford to make one of these with no job, and no programing skills, I will definitely ask your help.

    Just out of curiosity, but since I don’t have time right now to read the entire 70+ page instructable, does this thing account for the thickness of the spine/page deformation or text cut-off because of bulging spine shape? And how hard would it be to add a page turner to it?

    specifically, I have one book that’s a real problem- it’s a copy of Reese’s Cyclopedia from the 1900s on watchmaking, and very very densely illustrated, with old style text that’s very fine and close. It’s hard to read even with the book in hand. that said, could this handle that insane level of text density and still have perfect resolution, preferably at large size?

  4. you don’t need programming skills — we already made the software.

    the v-shaped plastic platen presses the pages flat, that’s basically the only way we can do this with cameras. check step 13 for a photo illustration.

    a page turner would be truly difficult. and i wouldn’t trust most of my collection to a mechanical turner, honestly. you’d be surprised at how fast it really is to operate.

    essentially, capturing your encyclopedia is more of a camera resolution issue than anything. my guess is that the current system with 8MP cameras is probably not enough if the pages are very large. chances are very good that by the time you get to building one, we’ll have cameras with suitable resolution.

    finally, think about spending that few hundred bucks on paying some enterprising college student to build one of these and digitize your books. this could seriously be a cottage industry.

  5. I don’t think a page turner has to be difficult. You could use a light suction to lift a page and then it’s just a matter of how you want your suction arm to be positioned and move.

  6. Very ambitiousproject. Loking forward for more models. :-)

    I agree that you don’t really need a page turner, if you only want to copy one book then just a camera with out any machanics at all is enough. I copy a book at an avarage of 3s per spread.

    About page turner, you will get carpal syndrome if you scan more than a few books at a time. Maybe making it easier to turn pages, and lift the scanning lid. Perhaps using air to suck up the pages like the one in Germany[1], and some easier way to manager the “lid”..

    [1]http://www.youtube.com/watch?v=y16rNqnxj0U

  7. I think the page turner is one of those ideas that seems easy until you try to do it, honestly. Just look at all the human augmentation that goes into page turning — tacky goop/saliva for fingers, book position, edges, the consistency of skin, etc.

  8. I did this with a single camera, a single halogen lamp, a tripod, and no platen. It worked OK, and it was probably faster since I didn’t have to move anything between pages. A platen shaped like —v— would help a lot keeping the pages flat and pushing the spine apart while letting me continue to use a single camera. When I did my setup (no platen), I used a 5Mp point-and-shoot, and the resolution was just enough to be readable doing 2 pages at a time. Now that I have a 12Mp DSLR, it should be a lot more readable and maybe a bit quicker.

    79 steps is a little excessive for this relatively simple setup, and I was disappointed that it did not have automatic page turning.

  9. I’ve tried your method, with a single camera over the book. Not holding the pages flat yields ugly output — the words are warped. 5MP is not enough for textbooks, though it is plenty for novels.

    How did you keep the book open without having your fingers in the pictures? Many books won’t stay open under their own weight.

    Because I wanted anybody to be able to make the scanner, I opted for verbosity and more pictures and explanation rather than fewer pictures. I wish more people would do the same with their projects. It is a lot more readable if you are using their generated PDF instead of the website.

    I am going to make a video comparing the speed of the automatic page turners to my scanner. It is not that much greater. And on my scanner, you can watch a movie while you work. ;)

  10. Haha! Is the quality so consistent that you can really watch a movie, then it’s really impressive.. :-)

    Yeah. With one 8MP camera in your hands you will get fingers in the pictures, and the pages will be warped. For smaller books under 200 pages it’s ok almost good quality, but as you say textbooks are harder. For me 8MP works fine, it’s very readable even though it’s not a perfect quality.

    Hmm what I really want is a portable ninja book scanner, that might be why I accept crappy quality.. :-)

  11. I would not call this a hack perse either, more a construction tip, and it’s nothing new really either, and the text needs to be OCR’ed else you have nothing searchable, and OCR’ing takes time and needs sanity checks too for the times it fails dramatically.

    As for turning pages, if banks have automatic bill counters that means you can reliable separate papers heh.

    Talking of scanning, did you see those torture memo scans? Why do those kinds of documents always look like they were retrieved from a crashed plane that subsequently was underwater for a week? Is that how they handle papers in washington? Weird stuff, and then after the abuse they are scanned at a weird angle, you’d expect perfect crisp copies, perhaps you should sell this scan setup to the pentagon.

  12. A0 is harder because it’s BIG (1mx1m) since you have to build a different mount, and it takes too much room. But it’s also easier since it lays flat on the surface, so it’s really easy to get consistent white balance over the whole surface.

    To shoot 0.5mx0.5m I’m using two bi-peds (tri-pods) and one beam between them to mount two cameras and lights. You might need more than two depending on your camera resolution, I had rented two HQ 20MP cameras which cost me $100 but it was worth it, at 120 pixels/mm² you can see some details.. :-)

    Best place to go is to the forums at: http://diybookscanner.org/

Leave a Reply to barry99705Cancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.