Simple, easy to build, ~600pg/hr, very portable scanner

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.

Moderator: peterZ

User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Simple, easy to build, ~600pg/hr, very portable scanner

Post by Mohib »

Hi everyone,

I'm new here and thought I'd just share my DIY book scanner design now that it's complete and working pretty much the way I want.

My design criteria were:

1) I wanted a totally knock down scanner, portable enough to carry on a bus so I could take it anywhere, clamp it on the end of any small table and go to work.

2) I wanted to be able to do paperbacks just as fast (or nearly as fast) as hardbacks, with as good quality and without damaging the spine by forcing the books open wide. I also wanted to handle sizes up to newspapers. The design can handle books up to A4 under the platen or, without the platen up to tabloid/A3 but at those sizes the original needs to be an unfolded photocopy so it's flat.

3) I wanted to be able to scan a fairly high speeds but didn't need ultra high speed that you get from a two camera set-up. Also two cameras would increase construction complexity and make a portable design harder.

4) It had to be built with an absolute minimal amount of handiwork with only whatever bits and pieces I could find at local hardware stores and just a few basic tools (files, power+hand drill, saw, vice). Also it had to be very solid with no finicky parts that need constant attention, adjusting and fiddling during operation.

5) It had to work with small, compact cameras without using electronic triggers and without using a computer so I use Daniel Reetz' bike break lever trigger (that idea was inspired thinking by Reetz).

I noticed the other designs -- even one camera systems -- almost always used some kind of V or other shaped frame/support to support the book. But, for me, this just created a huge, not very portable structure that also increased construction complexity. I decided that I'd use the table as the "frame" to support the book and to "connect" the parts together. So I only needed a way to press the page flat on the table and a way to support the camera over the book.

Since this had to be knock down, no glue was used between the pipe fittings. The bungee cord is hooked into an eye bolt which goes through pipe cover and attaches to the ball joint holding the camera and pulls the camera, and it's arm, tight against the vertical post. The T joint at the top of the main post is held in place with a hose clamp instead of glue so that it is solid and won't fly apart under the pressure of the bungee cord (which is another reason glue wasn't relied on the camera arm). And hidden inside the white, platen handle is a steel cable that's wrapped around the dowel nuts, which when tightened by the bolts going through the platen, pull the cable tight and holds all the handle parts together. There was generally enough friction between the platen handle pipe fittings to hold the weight of the platen while tilting during operation, but it does loosen after a while and this eliminated any accidental dismantling (which happened occasionally).

As the last picture shows, the whole scanner be knocked down for transport into 4 main parts: a) vertical post and clamp, b) camera + ball joint + camera arm (separate in the picture), c) platen handle + break lever (two bolts attach the platen to the white handle) and d) the platen (wrapped in the white towel). To knock it down, just release the bungee cord to release the camera and undo the two bolts platen handle bolts.

The book rests on a black anti-skid mat, and that rests on a plastic place-mat. I simply centre the book on the black mat and then just slide place-mat under it to centre the book in the view finder. My camera displays grid lines in the monitor which helps in getting the page straight. I also adjust the ball head to get the camera level so pages are not distorted due to perspective (ideally I need a small round spirit level to place on the view finder to get it properly square). The camera can also auto focus on any part of the image which is handy with pages like those at the end of a chapter where it's mostly blank, except for a para or 2 at the top.

To operate (as per the attached video), I simply hold the white handle with the break leaver in my right hand, tilt the platen towards the main support post, turn the page with my left hand and then hold the left side of the book up so the spine is not opened up wide, lower the platen and pull the break leaver to trigger the camera (you can see the brake cable coming up the side of the camera, and in the close ups you can see the flex in the bolt when the shutter is triggered). The rubber band across the platen handle prevents pages from flopping down over the platen or paperbacks from closing. Once I've scanned the odd pages, turn the book around and to the same for the even pages.

I'm able to do hard-cover books at 600 pages per hour. In the video in the attached zip file, after the first page where I demonstrate and explain the operation, I scan 7 more in 40 seconds at a leisurely pace. That's about 10 a min or 600 pages an hour but net throughput can drop to 200-400 pages an hour because of time wasted on separating stuck pages, weariness, the book shifting, being over cautious with the trigger, fumbling while turning pages, etc. It's clear from the video, the camera's auto-focus mechanism, not turning the pages, is using 65%-75% the time it takes to do a page and so I could probably increase the speed by at least 50% if I could lock my camera's focus and didn't have to pull the break leaver slowly and carefully to just activate the auto-focus and not trip the shutter and then pause while the camera focuses before pulling it the rest of the way to trip the shutter.

Once I've finished scanning I use Bulk Rename Utility to rename the files so the odd and even interleave correctly and then process the images with Scan Tailor (what a brilliant piece of software, which would be almost perfect if you could drag and drop pages to correct any out of order -- as sometimes you realise too late). I just use ambient lighting as colour balance, reflections, etc. are not too critical for me because my main objective was to create images for accurate OCR, but if colour is important you can add lights, nevertheless, at night I use a small incandescent bulb on a clip fixture, off to the side, and this also eliminates all reflections.

One problem with thick books is the distance to the camera changes from the start to the end of the book, so pages get smaller. This can be somewhat dealt with if your camera allows fine control of the zoom and you adjust it every 20-30 pages, but since mine (and most cheap cameras) doesn't have that kind of control, the ball head I have is a little odd in that it is mounted on the end of short, hinged extension that's about 1.5" long and lets me vary the camera distance up and down by 3". So if I have a very thick book I could keep the page sizes the same by making an adjustment every so often, but have not tried this yet as for me this was a minor issue given that everything else works so well.

Cost for the scanner was about $80 -- all the bits and pieces are from local hardware stores. Speciality items ordered (or sourced locally) which are $60 of the total are:

a) Manfrotto SuperClamp -- ($30 from Amazon). It's the most solid, flex-free clamp around and has a unique notch which let's the post be fastened solid to it. Also I can use with it the ball head for general photography when I don't want to cart a tripod around

b) Light-weight ball head -- ($10 from Amazon). Attaches the camera to the support and lets me level the camera. What's really needed is a small circular spirit level resting on it to get it spot on so there is no distortion.

c) Used bike brake leaver and cable -- ($10 from local bike repair shop).

d) Platen: 11"x15", 3/8" Plexiglas -- ($10, cut to size from a local plastics supply). I used 3/8" because it's thick enough so it doesn't flex, heavy enough to flatten pages with minimal pressure, cheap, won't break and let me drill holes in it with ease. It scratches easier than glass so although I keep it wrapped in a thick towel to protect it, it does get hairline scratches from rubbing against the paper, although you can't notice them. At $10 it's a cheap enough to replace.

e) Bolts/dowel nuts that attach the platen to the white handle from Ikea after a lot of begging till they parted with 2!
2014-03-24 17.51.44-small.jpg
2014-03-24 17.43.37-small.jpg
2014-03-24 17.38.38-small.jpg
2014-03-24 17.41.42-2-small.jpg
2014-03-24 17.44.34-small.jpg
2014-03-24 17.54.21-small.jpg
2014-03-24 17.53.38-2-small.jpg
2014-04-06 18.12.06-small.jpg
Attachments
Video of Mohib's DIY book scanner in action.zip
(5.74 MiB) Downloaded 1257 times
User avatar
davidlandin
Posts: 134
Joined: 22 Feb 2012, 15:44
E-book readers owned: kindle
Number of books owned: 0
Country: UK

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by davidlandin »

Very nice design. Can you explain why you have the two layers of material under the book? I think the black material is the non-slip drawer liner - is that right? What is the white stuff?

Have you considered infrared remote triggering of the camera?

David Landin
User avatar
daniel_reetz
Posts: 2812
Joined: 03 Jun 2009, 13:56
E-book readers owned: Used to have a PRS-500
Number of books owned: 600
Country: United States
Contact:

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by daniel_reetz »

Mohib, this is really nice work. Thank you for the extensive documentation, beautiful, informative photos, and most especially for taking the time to explain your design rationale at length. I love the way you integrated the bike brake handle with the platen - after watching your video I can really see how it speeds up your scan times.

A few thoughts - it seems like you could go even faster if the "far" side of the platen were secured by a compliant hinge. Rather than making something fixed or actually hinged, you might be able to get the same benefit by using a sticky mat draped on the platen plastic along the "far" side, and then touching the table. Kinda hard to describe, but it would adhere the end of the platen to the table, creating a flexible hinge. Another thought would again be to use something like the sticky mat to hold a stop in place on the table that the handle size of the scanner could rotate against.
dashmat.jpg
dashmat.jpg (42.02 KiB) Viewed 25360 times
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by Mohib »

Thanks David. Yes there are two plastic layers. One is a non-slip drawer-liner and it rests on top of a large plastic place mat. I found the second mat lying around the home and it's smooth on one side and dimpled on the other and looks like it's been cut off a roll, but any large stiff place-mat for tables would do. Everything rests on that and that rests on the able so I can slide the the book and platen under the camera to centre it in the view finder and get the page square. If you download the video, you'll see what I mean.

No, I haven't thought about infra-red triggers because I wanted a system that would work with any small, cheap camera (and mine doesn't have such a feature). Also unless you can see how to integrate the infra-red trigger with the platen handle so it's as smooth and efficient to operate as the bike lever I think it would slow the process down. The bike break is really just ideal ergonomically and very efficient, although I can see how an infra-red trigger would simplify construction and eliminate all the hardware around the camera.
Last edited by Mohib on 07 Apr 2014, 11:56, edited 4 times in total.
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by Mohib »

Thanks Daniel.

Yeah the bike break is really the heart of the entire scanner. Without it, it just would not be as efficient and smooth as it is.

WRT your other thoughts by "far side," I'm assuming you mean the side of the platen nearest the vertical post. If so, then I think I understand what you are saying and was thinking something similar. My thoughts were to do away with the black non-slip liner altogether and let the book slide on the table (or a slip as opposed to non-slip surface) so that it mimics the book slider typical V shaped scanners have. The platen would then need to have some kind of registration mechanism that you adjust once the platen is centered in the camera and then you fasten it to lock it so the registration mechanism doesn't move. This now ensures the platen is in the same place, relative to the camera, each time it comes down, and you can slide the book against it to get the book in position. The mechanism would need to register the platen both horizontally and vertically to be best.

One thought I had for this were two more holes in the platen in the other two corners and a steel cable running through them and to some kind of clamp attached to the Manfrotto Superclamp, then you adjust the platen and tighten this clamp to fix the cables tight (similar to the way the bike break cable is fastened to the trigger lever) so their length is fixed. This now holds the platen in a fixed x-y position relative to the camera and you just use the whole scanner as before but without the non-slip mat and simply push the platen away from the post as you drop it down so the cables are taught and engage the spine of the book, sliding the book left as you come down. Of course you just need to make sure the book is not moving vertically out of the frame.

I'm sure there are other ideas, like your very nice idea to simply make the non-slip mat larger so the the platen presses against it when it's lifted so it doesn't slide. Actually thinking about it some more, that's a really good idea! Simple and no extra parts to make and fiddle with. But overall it really does work fine because as long as the book doesn't move it's really effortless to slide the platen, while not letting go of the platen handle, so the left edge engages in the book spine.
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by Mohib »

BTW I'm not sure if the author of ScanTailor reads these forums, or perhaps someone here is in touch with him, but one very, very useful improvement would be a new step between steps 4 (Select content) and 5 (Margins) which I would call "4A - Content size normalization".

What you do, is select a typical page of text, and the width of that content becomes the baseline. You then say "apply to all pages", "selected pages", etc. to scale the other pages' content so the width matches the baseline page, and then the margins are applied.

This way problems with varying page sizes on images -- and hence content size (i.e. text is not uniform size on every page), due to camera-book distance slippage as you go through the book (as I described I have, and other simpler scanner designs have also, although not a problem in the typical v-platen, fixed camera position designs) can be eliminated during post processing and the content -- i.e. text size, and line lengths -- will have uniform size. I'm sure there are other details that need to be thought through, but that's the basic idea.
User avatar
davidlandin
Posts: 134
Joined: 22 Feb 2012, 15:44
E-book readers owned: kindle
Number of books owned: 0
Country: UK

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by davidlandin »

Mohib wrote:BTW I'm not sure if the author of ScanTailor reads these forums, or perhaps someone here is in touch with him, but one very, very useful improvement would be a new step between steps 4 (Select content) and 5 (Margins) which I would call "4A - Content size normalization".

What you do, is select a typical page of text, and the width of that content becomes the baseline. You then say "apply to all pages", "selected pages", etc. to scale the other pages' content so the width matches the baseline page, and then the margins are applied.
I think you can do what you want if you use YASW (Yet Another Scanning Wizard), which I've made a YouTube video about

http://youtu.be/__a9urAtQD4

I have used YASW to do stage one post-processing, then Scan Tailor afterwards.

I'm still not clear why you need TWO mats under the book! Excuse my ignorance. . . .

David Landin
User avatar
Mohib
Posts: 107
Joined: 05 Apr 2014, 21:15
Number of books owned: 0
Country: Canada

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by Mohib »

Hi Thanks for the video for YASW. Interesting product, but it won't do the scaling I'm referring to.

Here's the problem with a thick book. Say we're doing a 1,000 page book. I do the right hand (odd) pages. Page 1 is closer to the camera and so is imaged larger. By the time I reach page 999, the "pile of pages" as disappeared and so page 999 is farther from the camera and so is imaged smaller.

Now to do the left hand (even) pages quickly I spin the book around and start at page 1,000 (which doesn't have a facing page) so I'm turning the pages "normally" and can work fast. But since page 1,000 is at the "top" of the pile of pages it gets image larger. Page 998, the next page to be scanned is also large, while it's facing page, page 999, was imaged small. By the time I reach page 2, it's again at the bottom of the pile of pages and so is imaged small, while it's facing page, page 3, was imaged large.

I've attached an example so you can see he difference it made in just a 230 page book. By the time you convert this to a PDF and are flipping through, the pages are jumping back and forth between large and small (and of course wide and narrow) and although liveable, it's actually a bit annoying.
A Short History of the Ismailis - p 010 v p 011 - 35%.jpg
A Short History of the Ismailis - p 010 v p 011 - 35%.jpg (390.28 KiB) Viewed 25315 times
Now this scaling is gradual and varying continuously, ever so slightly, across the whole book, page by page. So what I'm suggesting is that after ST selects the contents (step 4), and before it applies the margins (step 5), you select a page that you'd like to be the baseline, and it scales the selected content on every page (or selected pages, etc.) to match that size (in width because height may vary as content may not fill the page). There are other ways you can do it -- pick the first and last pages (assuming they are the largest and smallest) from one side of the book, calculate the difference and determine the adjustment factor to apply to every page. And so forth. I'm not sure which would be the most fool-proof and reliable, and some experimenting would be in order.

The benefit of this is that simpler scanner designs, where the platen to camera distance is not fixed, can have their results tuned up in post processing to have the high quality consistency of the more sophisticated V platen scanner designs, with their fixed platen to camera distances. And the better the quality that can be produced from simpler scanner designs the more such scanners will be built (as many can't make the complex ones) and the better for everyone.

As for why I need two mats, I think if you're able to see the video you'll see. The book is on the non-slip mat so it doesn't move while scanning (and also provides the black background needed), but that is hard to slide across the table (as it's a non-slip mat) when centering the book in the camera, or getting the page square, so it rests on another mat which is easy to slide. Just using one or the other doesn't work so well I found. But the combination and weight of the book and platen seems to work fine.
User avatar
daniel_reetz
Posts: 2812
Joined: 03 Jun 2009, 13:56
E-book readers owned: Used to have a PRS-500
Number of books owned: 600
Country: United States
Contact:

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by daniel_reetz »

Mohib wrote:But overall it really does work fine because as long as the book doesn't move it's really effortless to slide the platen, while not letting go of the platen handle, so the left edge engages in the book spine.
Yeah, this was totally obvious to me when I saw the video. Clean and efficient. Again, great work.
TomHorsley
Posts: 96
Joined: 30 Jan 2011, 10:39

Re: Simple, easy to build, ~600pg/hr, very portable scanner

Post by TomHorsley »

It would complicate a wonderfully simple design, but you could eliminate the scaling change from page to page if you could fix the platen distance from the camera and press the book up from the bottom against the glass. I keep thinking about a design like that, but haven't managed to work out any details that seem satisfactory in my imagination, much less in reality :-).
Post Reply