Minimum Camera for Book Scanning

[roger.carden]

What is the minimum specs that I should look for in a camera for Book Scanning?

[daniel_reetz]

This question gets asked very often. I think we need a "FAQ" on the topic. Roger, I strongly recommend searching the forums, as the recommendations made for others really haven't changed. I still think the Powershot A590 IS is the best camera for the job in terms of price and performance. Please have a look at the other build threads to see what people are using.

Someone recently asked me this question via email, and here is what I wrote.

We use these criteria when selecting cameras:

1. SDM/CHDK capability (for our USB triggers).

2. Resolution higher than 8mpix (mostly important for larger texts. If
your book is small, 7 is acceptable. When pages approach A4 you need
more resolution).

3. Some manual control available via the stock firmware or CHDK.

Power supplies are available on eBay for almost every Powershot model
for around $12. I still use batteries, but my next scanner will use
these cheap power supplies.

AFAIK CHDK and SDM do not work with the A480, so it doesn't work for
the cheap USB triggers that emit a 5V pulse. Canon has also been
phasing out software control for these cameras. You will want to check
with Canon or the Gphoto2 people about whether or not it is supported,
or trigger them mechanically (but again, you might not have enough
manual control to do a good job).

The G series canons are good but expensive.

The A590 is still a favorite, often available refurbished for $120.
Watch for Canon Outlet specials or put in a Google Alert for
"refurbished A590 IS".

AFAIK there is no alternative to CHDK other than mechanical
triggering, which is possible. One of our forum members is working on
a system using bicycle brake levers. The 790 IS look like good cameras
to me. If you want more discussion on the matter, I recommend getting
an account at DIYbookscanner.org/forum or at least searching the forum
there. I prefer to keep this kind of communication in the open so
everyone can benefit. You can also see many users there using
non-canon cameras, DSLRs, etc. Many people have asked this question in
the forums and there has been a lot of discussion around it.

[dtic]

How important is it to get two cameras of the same model? Would for example a powershot a710 + ixus 70 (both 7 megapixel) work about as well as two 710's?

[daniel_reetz]

You'll have to deal with each set of images a little differently, but it probably doesn't matter much as long as they're not radically different resolution. Spamsickle had different cameras on his rig forever - I think an SLR and an Ixus at first. If you check out his posts, you can read more. And though it is slower, you can always start with just one camera.

[Misty]

If you're scanning mostly-text books to go into Scan Tailor, you're in an especially good situation for unmatched cameras - the image itself is going to undergo such heavy processing that any differences in the cameras' images will be pretty much unnoticeable once you get the Scan Tailor output.

[Antoha-spb]

And though it is slower, you can always start with just one camera.

=== it is NOT much slower with the properly installed platen



and will even save some time on combining pages from two cameras.....

[Zenock]

or trigger them mechanically

So assuming that I'm doing this, what other things should I be considering.

I mean if I just buy the cheapest 8 MP camera I can find, is that going to work? What am I missing, what do I want to watch out for?

Thanks,
Z.

[daniel_reetz]

The Canon Powershots have been proven, image-quality wise, to be good enough. If you can find another cam that seems good enough, there's no reason not to buy it, really. Mechanical triggering works with every single camera ever, because mechanical triggering is just button pushing. For purchasing criteria, I'd look for samples of images online from whatever cam you are interested in, and I'd look for cameras which look easy to attach stuff to, to make the mechanical part easy.

[gouthro]

Hello,
I know practically nothing about this subject and so my question may seem silly. But, I have taken a look at the various posts and can't find anything close to what i need. i don't even know what kind of question to ask to look for a previous thread. But, here goes: I just want to use my camera to scan not books, but to take notes. Say, a page or two here or there. But, I need to have OCR and a picture from my camera doesn"t seem to do that. What am I missing? I have an old camera ( 2001) and an old version of Microsoft Word (2003). Can I do get OCR with this setup?

thanks gouthro

[Misty]

You need to process the page from your camera using an OCR program. The OCR program will take the JPEG and turn it into text that you can put into Word. There are a few free open-source OCR programs like Tesseract (accessible via an online interface here), and commercial ones like ABBYY, Omnipage, Readiris, etc.