Lego Mindstorms scanner - A novel way of scanning magazines

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

JasonStonier
Posts: 19
Joined: 16 Nov 2021, 06:29
E-book readers owned: Kindle Paperwhite
Number of books owned: 2000
Country: UK

Lego Mindstorms scanner - A novel way of scanning magazines

Post by JasonStonier »

I have something like 2,000 old New Scientist magasines which I want to convert into a digital archive - pre-1989 the majority of them are already in the public domain, and post 2007 they're available as digital downloads but there's that period of the 1990s and early 2000s where all I have is the hard copy.

And also, I like technical challenges - so I set out to build an automatic book scanner out of Lego, because...why not.

https://www.youtube.com/watch?v=v15JGP2vJ_I

This is the Gen-1 device - it works surprisingly well. An 80mm PC fan is used as a vacuum to pick up the page, a 20mm PC fan is used to pick up one corner to reduce the "sticktion" between pages, and a third fan introduces flutter to the page edge to improve the chance of only picking up one page.

Once the page is turned, a lego motor clips a microswitch which is wired into two bluetooth controllers for the cameras on the Android phones. Images are automatically uploaded to a linux server into directories for LEFT and RIGHT which I then process with ImageMagik into PDFs. Occasionally it will turn two pages at once - that's about 1 in every 100 turns, and more often it will get stuck on a page and not be able to turn it - that's about 1 in 50 pages.

Only problem with it is it's very slow - about 20 mins per magasine, so I'm working on a Gen-2 version which uses a Czur Shine overhead scanner and I hope to get the workflow down to about 10 minutes per magasine.
cday
Posts: 447
Joined: 19 Mar 2013, 14:55
Number of books owned: 0
Country: UK

Re: Lego Mindstorms scanner

Post by cday »

Ingenious and I'm pretty sure original! :D

A method I used to scan magazines was to destaple them, and then feed the loose sheets through a sheet feed scanner. The public library here in Cheltenham has an excellent office-grade multifunction printer/copier/scanner which is available for public use, which can optionally scan both sides in one run. With four pages scanned per sheet at least 100 pages can be scanned in one pass, with luck without any misfeeds. The magazines I scanned were A5 format, so the maximum supported sheet feeder sheet length might limit use for significantly larger format sheets.

The pages then, of course, have to be cropped and renumbered, for which I developed a batch file that used XnView's NConvert for the image processing and CMD for the logic required. The algorithm required some clear thinking and plenty of cheap coffee in a quiet area of Wetherspoon's, I haven't got into scripting in Linux yet.

Reassembling the magazines, if required, is optional and there might be scope for some optimisation as a separate project there... ;)
JasonStonier
Posts: 19
Joined: 16 Nov 2021, 06:29
E-book readers owned: Kindle Paperwhite
Number of books owned: 2000
Country: UK

Re: Lego Mindstorms scanner

Post by JasonStonier »

I did try to de-staple the magasines, and run them through an A3 photocopier at work...but it consistently mangled the pages.

To be honest, the amount of time I've spent developing that page turner, I could have just scanned the magasines by hand...but it's at least partly about the challenge.
cday
Posts: 447
Joined: 19 Mar 2013, 14:55
Number of books owned: 0
Country: UK

Re: Lego Mindstorms scanner

Post by cday »

JasonStonier wrote: 16 Nov 2021, 10:32 I did try to de-staple the magasines, and run them through an A3 photocopier at work...but it consistently mangled the pages.
Through a sheet feeder? That seems bad luck, you were scanning double-sided, which might have been an adverse factor? I did have some trouble myself as the paper used in the magazines was unusually shiny and sometimes needed multiple attempts to scan, but most paper would have scanned without any real issues, I think. I think now that the library scanner I used would actually have scanned A3 loose sheets. And then you used an algorithm to name the cut pages in the correct sequence?

To be honest, the amount of time I've spent developing that page turner, I could have just scanned the magasines by hand...but it's at least partly about the challenge.
I quite understand... ;) I think some other forum members have in the past persisted with automated scanner projects against the odds!

There was an amazing fully-automated scanner design posted some years ago by a Japanese guy. It attracted very little attention at the time and unfortunately the text was in Japanese, and a possible translation by a forum member never materialised. I haven't been able to think of enough keywords to find it, not helped by the fact that the content consisted mainly of the video with Japanese commentary.
JasonStonier
Posts: 19
Joined: 16 Nov 2021, 06:29
E-book readers owned: Kindle Paperwhite
Number of books owned: 2000
Country: UK

Re: Lego Mindstorms scanner

Post by JasonStonier »

Yeah, it was odd with the sheet feeder - would run through a stack of normal printed documents just fine, but always mangled the magasines. I had to iron them in the sandwich toaster afterwards (true story - works brilliantly to flatten a magasine that has curled due to age or poor storage).

I'm partly through re-building the scanner in a completely different design which I hope will be quicker and more reliable - we'll see. Since doing that original design I have gotten into 3D printing and so I've made up some special Lego-compatible bits to improve the design somewhat.
cday
Posts: 447
Joined: 19 Mar 2013, 14:55
Number of books owned: 0
Country: UK

Re: Lego Mindstorms scanner

Post by cday »

I think on reflection that magazine paper is probably generally not very similar to typical office paper, so batch scanning loose sheets may more often encounter problems, even though for selected magazines it may be a viable and potentially very efficient way of archiving large collections.

So maybe your 'air jet non-contact' method of imaging is worthy of further research and development! :D

And perhaps your novel sandwich toaster method of flattening distorted pages might be worthy of a separate thread? ;)
JasonStonier
Posts: 19
Joined: 16 Nov 2021, 06:29
E-book readers owned: Kindle Paperwhite
Number of books owned: 2000
Country: UK

Re: Lego Mindstorms scanner - A novel way of scanning magazines

Post by JasonStonier »

The whole ironing magasines thing - I promise it's the best ever answer to your spouse asking "where the hell is the sandwich toaster".
Attachments
ToasterMagasine.jpg
cday
Posts: 447
Joined: 19 Mar 2013, 14:55
Number of books owned: 0
Country: UK

Re: Lego Mindstorms scanner - A novel way of scanning magazines

Post by cday »

JasonStonier wrote: 17 Nov 2021, 05:37 The whole ironing magasines thing - I promise it's the best ever answer to your spouse asking "where the hell is the sandwich toaster".
Yes, yes, but isn't worth its own thread to raise its visibility? :lol:
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: Lego Mindstorms scanner - A novel way of scanning magazines

Post by daniel_reetz »

This is really cool!

It really hits on one of my favorite concepts, which is that if a bookscanner can be made totally automated, it can go slowly - possibly very slowly - and still be really useful.

The sort of flattening/ironing concept you have is also really cool.
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: Lego Mindstorms scanner - A novel way of scanning magazines

Post by daniel_reetz »

JasonStonier wrote: 16 Nov 2021, 08:21 Only problem with it is it's very slow - about 20 mins per magasine, so I'm working on a Gen-2 version which uses a Czur Shine overhead scanner and I hope to get the workflow down to about 10 minutes per magasine.
Another way to solve this problem is to build 2 copies ;)
Post Reply