A Scanner for books with text VERY close to the gutter.

[JThomas]

I have a couple of books that I am looking to scan. The problem is that the text is VERY close to the gutter in most places.
Two pieces of glass, even at greater than 100 degrees just won't do it.

So I am building a scanner with just this in mind.
When I am done building it, I will post pics. It will only be able to do one page at a time so only one camera is required.
This might be a good project for someone on a budget.

I will be using a Canon XS with a remote switch for taking the pictures.

You will have the ability to turn pages very fast. I think, fractionally faster then any I have seen on this site so far.
Time will tell on that one though.

I am real close to completion so stay tuned.

[JThomas]

Oh, I forgot. Here's a picture of one of the books I'm talking about.

This is a page I opened up to randomly. Some are better and some are worse.

[daniel_reetz]

Wow, that is rather tight. I assume you are using very thin glass? Maybe Gorilla glass would be a good option for you.

[JThomas]

I am using a thin piece of glass salvaged from a flatbed scanner. It's works great in some tests I've performed.

Just waiting for my paint to dry and then a little more work to make everything adjustable, and then I'll be done.

[JThomas]

Here's what I have so far, and as far as capturing the text close to the gutter it works perfect.
The pressure from the bottom edge of the glass serves to press the opposing page out of the
line of sight. (Second picture)

The kit lens that comes with the XS actually let's you get 2.5" closer then the manual says. I'm about 8"
from the glass with some extra zoom to spare. This allows me to get almost all the 10MP on the page.

The camera is on the pivot point so it takes virtually no time for any vibration to settle down. What I'll
do before I start taking a bunch of pictures is to take the first one with the auto focus, and then use some
blue painters tape to keep the focus in place then switch off the auto focus not to wear it out. I'll just put it
in manual mode and use the same shutter speed, aperture, and ISO as the first picture. I have an AC adapter
so the batteries are not a concern.

To move it up and down I can either grab it from the part extending out beyond the camera, or
right in front close to the glass. I will have to use it a little while to determine the best place for me.
Then I will mount the remote shutter switch there.

I still have a few details to work out, such as something to mount a light to which is very simple.

Another is a good way to hold the book, but I have a few ideas on that also.

You can also adjust the distance to the camera as well as moving the glass up and down to keep
all sizes of images centered to capture as much of the page as you can with the most MP's.

Any comments or suggestions?

000 (Medium).JPG (71.23 KiB) Viewed 3524 times

001 (Medium).JPG (69.11 KiB) Viewed 3524 times

002 (Medium).JPG (56.38 KiB) Viewed 3524 times

003 (Medium).JPG (66.71 KiB) Viewed 3524 times

[daniel_reetz]

Absolutely gorgeous design and execution. Do you mind sharing a single page showing how close you can get to the gutter?

Also, how are you dealing with glare from room lights?

[JThomas]

Here is a preliminary shot taken holding everything by hand, including the light.
The bottom edge of the picture is the actual edge of the glass. I actually have some room to spare on this page
but others are closer. I didn't look for the worst case.

You can see the reflection of the opposing page in the bottom corners which I will have to work on.
It shouldn't be a big problem since I'm going to OCR the entire book (2 books actually with a total of about 2700 pages)

The font is a early 1900's German Gothic so I will have to train the OCR for it.

As far as the room lights goes, they will be off when I do it for real. With this big a job I want everything perfect
before I start.

[Oxford eBooks]

That's not too bad, though I notice that in that shot you haven't quite got ALL the page to the gutter. Did you manage it in later trials?
That font looks like it'll be a little tricky to OCR - what software are you planning to use?

[JThomas]

Oxford-

Here is a final shot with all the bugs worked out. (cropped for posting). You can see the gutter and a little of the other page. I have the camera zoomed
to bottom edge of the glass precisely.

I am using ABBYY Finereader for OCR. I am training it specifically for this font. Seems to be working great so far.

When using their default OCR it takes into account all kinds of fonts so will not be that precise as opposed to training it for
only one font.

Daniel: Maybe you can add this to your scanner build index.

[Oxford eBooks]

That's a really clean scan!
Yes, Abbyy is a great OCR, pity it tends to fall down with quotes even on simple fonts though.