Daniel Reetz, the founder of the DIY Book Scanner community, has recently started making videos of prototyping and shop tips. If you are tinkering with a book scanner (or any other project) in your home shop, these tips will come in handy. https://www.youtube.com/channel/UCn0gq8 ... g_8K1nfInQ

Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Built a scanner? Started to build a scanner? Record your progress here. Doesn't need to be a whole scanner - triggers and other parts are fine. Commercial scanners are fine too.
Post Reply
jesu_krist
Posts: 5
Joined: 10 Sep 2017, 08:31
Number of books owned: 0
Country: Italy

Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Post by jesu_krist » 11 Sep 2017, 13:57

Dear Users,

I've been following the feats of the "DIY Book Scanner" community for years, and now -- as a new user -- I'm glad to be in a position to share my simple but efficient scanning rig with you. I admire the fantastic work that is showcased in this forum, but I don't have access to the tools and hardware that are necessary for such big projects, so I opted for the simplest design that could fit my needs, i.e. a cube of plexiglass of 33 x 33 x 33 cm with an open face.

Image

Image

The two cameras (wall powered Canon PowerShot A2500) are inside the cube, both secured with two inexpensive suction cups, each facing toward the respective page of the book that sits open at a 90° angle; the ruler helps to center the book with reference to the cameras, and is very practical as a "stopper" in the fast pacing workflow of mine, as the book slides easily in place. The transparent plexiglass makes lighting very easy: I light everything from the sides, without any glare in the page area of the acquired pictures.

For the software part, I use the TwoCamControl AutoHotKey script with a little twist: previously I used a secondary keyboard as a "foot pedal" with all the keys re-mapped to F8, and this was achieved with the help of LuaMacros; now I voice control the "Shoot" function of TwoCamControl using Vocola 3 coupled with the speech recognition utility of Windows (I've recently switched to Windows 10 but this feature is also available for previous versions of Windows); now I'm able to trigger keystrokes with my voice, and this has really increased my productivity: now I'm able to capture more than a 1000 pages per hour with very little effort. The post-production is done with ScanTailor, with optimal results for black and white output.

The scanner is in daily use, and is serving me very well; let me know what you think about it.

This should exemplify the quality of the acquisition and the output of the post-production:

Image

Image

User avatar
jck57
Posts: 375
Joined: 23 Nov 2009, 15:21

Re: Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Post by jck57 » 12 Sep 2017, 10:23

Simply brilliant. I would love to see details of your lighting set-up and a video of your scanning operation. Thanks.

dpc
Posts: 251
Joined: 01 Apr 2011, 18:05
Number of books owned: 0
Location: Issaquah, WA

Re: Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Post by dpc » 12 Sep 2017, 13:17

Thanks for sharing your design. Good job!

I had built something like this a few years ago except it was a single camera scanner. I did it to capture sample pages while I was developing my software post-processing pipeline so the throughput wasn't that critical. Imagine an open-sided retangular tube that had a piece of clear acrylic sheet on one end with two lights on left and right shining in from both sides at the platen surface. Took about an hour to build, was lightweight and easy to store, and was pretty helpful in acquiring images from various types of books from my library.

The takeaways from this experiment:

1) Flipping the book over to turn the pages was cumbersome. I wouldn't shoot all of the odd pages, then restart from the beginning and shoot the left side pages. I would shoot a page, rotating the book and shoot the adjacent page, then turn the page. I did it this way because I initially thought page-turning was a source of errors (stuck pages, skipping pages, etc. - more on this is mentioned below), and it can put more stress on the book than only having to turn the pages once.

2) The acrylic sheet attracts dust and small paper particles moreso than glass and those particles become difficult to keep off the plastic surface during the scan. I chamfered the edge of the acrylic that goes into the book's gutter so that it wouldn't rake across the page when I flipped it each time to shoot the next page. Raking that sharp edge would generate additional paper particles.

3) The constant sliding of the book on and off of the acrylic can cause scratches on the platen surface that maybe visible in your images. This isn't that big of a deal if you are post-processing with ScanTailor to generate black/white images containing only text. If you're scanning something with a darker-colored image (magazines, books of photographs, etc.), you can see these artifiacts quite easily. Moving to a glass platen would help a lot here.

4) I had a problem skipping pages. This is more a problem, I believe, of me rotating the book when shooting the left, then right side pages because I'd have to alternate between rotating the book and rotating the book combined with turning a page. You obviously won't have this problem with your dual page scanner. Sometimes I would forget after flipping the book if I was supposed to rotate and turn the page or simple rotate it. My DSLR was controlled by an app I wrote on my laptop and I would ensure that the file name of the jpg file (based on page number) matched the page number I was currently scanning. Every chapter or so I'd check that those two numbers matched to determine if I had skipped any pages along the way.

5) I would mask off the portion of the platen surface that wasn't being covered by the book with black paper. While the room was darkened when I would scan books, the scanners dual lights which were pointed at the platen surface would shine through the portions not covered by the book and would illuminate whatever was in the background. This was easier than trying to hang a black sheet in the background as I may not be scanning books in the same location.

6) I would get slight movement of the scanner when using it. This posed a problem because in my post-processing stages one of the things I do is perform uneven lighting correction based of a shot of a 50% gray card before I start scanning a book. Since my lights were not fixed relative to the platen surface, any movement of the lights or scanner would invalidate the lighting calibration image. I ended up clamping things down to the table top and that solved that problem. Again, that's probably not an issue for you.

One potential problem that I can see with your scanner design is that when you remove the book from the platen to turn the page, the person operating the scanner is going to have a bright light shining up at them and into their eyes. In a darkened room I can imagine that this may be very distracting. Have you found that to be an issue when scanning books using your scanner?

Any problems with reflections from the adjacent plexiglas surfaces?

jesu_krist
Posts: 5
Joined: 10 Sep 2017, 08:31
Number of books owned: 0
Country: Italy

Re: Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Post by jesu_krist » 20 Sep 2017, 09:48

This is an update on the scanning rig:

Image

Image

The base is just a cheap white shelf (60 cm x 40 cm); the cube is held in position (at a 45° angle) with four plastic shelf brackets; the rest is cable management. For lighting I'm now using a little LED tube light (reading straight from the box: LED, 4W = 34W, 380 LUMEN, 4000K), and this has removed the subtle glare and reflection from the adjacent plexiglas surfaces that I was experiencing in the page area; these are the new inputs and outputs:

Image

Image

Answering to dpc, there is some light shining up, but with such a little LED light it is not annoying at all. The base also makes the design quite portable. Soon I'm going to post a video of my scanning operation (thanks to jck57 for the suggestion).

N.B. I'm aware that there is a little bit of "keystoning" on the input image.

recaptcha
Posts: 48
Joined: 03 Sep 2010, 13:23
Number of books owned: 0
Location: Calgary, Alberta, Canada

Re: Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Post by recaptcha » 20 Sep 2017, 13:39

Interesting Jesu_Krist.

1. Did you buy the plexiglass already made as a cube? Or did you buy the plexiglass pieces and assemble it into a cube? If the latter, how did you fasten the pieces together?

2. Theoretically, it seems the bottom half of the cube that holds the cameras does not even really need to be plexiglass, it could just be some other non-transparent material, like wood. Is this correct?

3. Do you find you have to move and adjust the position of the cameras for each new book size ?

Great job with the improved lighting! Looks nice.

jesu_krist
Posts: 5
Joined: 10 Sep 2017, 08:31
Number of books owned: 0
Country: Italy

Re: Two cameras in a cube of plexiglass (inverted design, ≈ 1000 pp/h)

Post by jesu_krist » 20 Sep 2017, 22:00

I'm going to answer each specific point:

1) This is my first serious attempt at a reliable and heavy duty scanner, so I just ordered the cube online, it was made by a shop specialized in plastic materials; the cost of the assembly in itself was very limited, and I had no shipping costs (the shop was in my city). This decision was mainly out of fear of screwing things up, but it is totally feasible to buy plexiglass and glue the pieces together: my local hardware store sells plastic materials, and cuts them for free; there I can buy specific glues to fasten the pieces myself.

2) The reason why I've chosen an all plexiglass build is because of the mounting solution for the cameras: I'm using suction cups. I'm not aware of any cheap material that can be easily glued to plexiglass (N.B. the aforementioned special glues) and that also works well with suction cups: of course you can use some other plastic material, even better if non transparent; for me it was easier of do the whole thing out of the same plexiglass.

3) I'm never adjusting the position of the cameras: I'm able to capture pages up to 28 cm H x 18 cm L, or even more if I remove the ruler; this is good enough for most book sizes. A bigger book would be to too heavy and impractical for this kind of inverted design.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest