Platen: Theory And Practice

DIY Book Scanner Skunk Works. Share your crazy ideas and novel approaches. Home of the "3D structure of a book" thread.

Moderator: peterZ

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:

Platen: Theory And Practice

Post by daniel_reetz »

This is the beginning of a series of posts going over platen issues. I am going to try to re-discover, from first principles, the ideal relationship between the platen, camera, and lights. My reason for doing this? I'm much better with optics than when I started DIY Book Scanner - so it's time to go discover WHY all this stuff worked so well from the beginning, and to see if there are any new insights to be gained into why things fail the way they do.

The first video deals with the "overhead view" of the platen. What does a light or camera "see" when directly overhead the platen, and how does that affect our choice of platen angle?


The second view deals with the "normal view" of the platen. The "normal" view is the camera looking directly at one glass face of the platen. How can we adjust the platen to work best for the camera?


Third and probably final platen video showing how the position of the lights causes or prevents glare images:
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: Platen: Theory And Practice

Post by daniel_reetz »

I would greatly appreciate any comments, criticisms or insights on these videos. Thanks.
Turtle91

Re: Platen: Theory And Practice

Post by Turtle91 »

Thanks Daniel!
I was interested in the first video when you put it up earlier, but didn't quite know how it would effect my design process.

The second video did a great job showing how the platten angle can effect the reflection onto the other pane. I was dreaming about this exact issue in the early morning drowze before actually waking up this morning (can you tell I'm deep into building my first scanner!) and was thinking about angles of incidence and reflection. In a perfect world a 90 degree platten wouldn't reflect any coherent light into the camera if the light was directly overhead and the camera was outside of a vertical line going up from the near edge of the platen. (I'll add a diagram later to better visualise), but you have the camera optical axis centered on the platen which messes up my perfect theoretical world!! I'm thinking that the distance of the camera from the edge of the platen will change the platen angle required to hide the reflection. ie. If the optical axis from your camera is further from the edge (larger pane) then the angle of the platen will need to be larger. Conversly a smaller platten would need a smaller angle (down to 90 minimum).

The third video also helped alot for someone like me who hasn't seen all these lighting issues before - simply read about them on this forum. I wonder if there is a basic rule-of-thumb that says how high the lights need to be for a given platten size and camera distance?? That would be nice when designing the height requirement for the light pole. I guess I could just set up the whole scanner and test it out before cutting the wood for the light post... But I really like those equations...there's got to be one.... x= 3y^2 + sqrt(pi) or something... lol

Cheers!
the.traveller
Posts: 82
Joined: 22 Sep 2010, 03:58
E-book readers owned: Samsung Tab S
Number of books owned: 800
Country: Netherlands
Location: Rotterdam, Netherlands

Re: Platen: Theory And Practice

Post by the.traveller »

Turtle91 wrote: (I'll add a diagram later to better visualise), but you have the camera optical axis centered on the platen which messes up my perfect theoretical world!! I'm thinking that the distance of the camera from the edge of the platen will change the platen angle required to hide the reflection. ie. If the optical axis from your camera is further from the edge (larger pane) then the angle of the platen will need to be larger. Conversly a smaller platten would need a smaller angle (down to 90 minimum).

The third video also helped alot for someone like me who hasn't seen all these lighting issues before - simply read about them on this forum. I wonder if there is a basic rule-of-thumb that says how high the lights need to be for a given platten size and camera distance?? That would be nice when designing the height requirement for the light pole. I guess I could just set up the whole scanner and test it out before cutting the wood for the light post... But I really like those equations...there's got to be one.... x= 3y^2 + sqrt(pi) or something... lol

Cheers!
Lucky you, I have been calculating those kind of things, together with the help from some other members.

Go to Science! Or, a new page on the wiki for technical works Post by rob » 01 Apr 2011 for a Excel sheet which has those calculations. Download it and start playing with the numbers.

Mind you I didn't incorporate the insights of these movies shown here. But you can change the platten angle to 100 in the sheet and it will give you the new size of your scanner and tells you how high your camera needs to be to make perpendicular pictures.

I still need to incorporate the calculations for the thickness of a book.

If you have time just read
Fix Keystoning via hardware adjustments Post by jay » 18 Apr 2011
and
Debate: camera positioning -- flexible vs. fixed Post by jerrytsai » 02 Jul 2011

There is one negative side of making the angle larger, that your camera needs to go further away (highered) from the surface of the platten. This if you do not want any keystoning etc happening. In this way you don't need to do a lot of preprocessing.
Look at Software Walkthrough, (strider1551's method) how strider has his workflow worked out.

I just thought of one positive thing about the higher camera position. Your lights now can be below your camera thus avoiding dropping shadows from the camera.
bnz

Re: Platen: Theory And Practice

Post by bnz »

Wow, these are some awesomely interesting videos! While I don't know enough about optics to contribute to this discussion, I have learnt a whole lot by watching them. Personally, if I ever build a second scanner, I would probably go for the 100-120 angle as the books that I scan aren't that old for it to matter or so and I'd rather not deal with the ghosting. My current scanner is 90 degrees, I guess it's a good compromise to live with with what I need right now...
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: Platen: Theory And Practice

Post by daniel_reetz »

Yeah, the 90 has the MAJOR advantage of being very easy to construct, which is why I chose it from the beginning. Nothing particularly wrong with going that way, but as always, we can improve... I'm not ready to recommend 100 degrees to everyone yet but I'm pretty close to saying it's ideal. Constructing a scanner with that angle now...
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:

All about glass.

Post by daniel_reetz »

I just bought glass for YET ANOTHER PLATEN and it's time to share what I know now.

1. "Doublestrength" glass, as it is sometimes sold in hardware stores, is just 1/8" glass. Nothing special about it, the name only specifies the thickness. IMO, this is the correct glass for book scanning. Thicker causes problems with color shift and image shift, as well as makes it tougher to get into the gutter of a book. Thinner, and you risk things breaking.

2. Putting a 45 degree angle on the edge of glass is called "Mitering" not "Beveling". "Beveling" is the angle you see on a mirror or something which does not continue down the edge of the glass. Mitering is cut all the way down to the far edge. In my area, nobody will try to miter or bevel glass that is thinner than 1/4", so for now my scanner will have unmitered glass.

3. Having the sharp edges of glass trimmed off is called "seaming". Seaming is a good idea, although I'm not yet sure if it is appropriate for the bottom-most seam.

4. Glass is actually cheaper, at least in California, from a glass dealer/frame shop than it is from a hardware store, and the quality is much better. At OSH hardware, a 24x36 sheet of doublestrength glass is 28.99 and they have to cut it for you and they do a Very Bad Job. I just got two pieces of glass professionally cut and seamed at El Cerrito Glass in El Cerrito, CA for $21.30.
Attachments
glass_2011-09-08_14-33-14_55.jpg
Mangan
Posts: 17
Joined: 19 Jan 2012, 14:33
E-book readers owned: Sony Xperia Arc
Number of books owned: 1000

Re: Platen: Theory And Practice

Post by Mangan »

About lightening the platen. What if having many small LED:s/lamps (4-6?) on each side of the platen. It looks like it would be quite evenly lit?
blueblazer

Re: Platen: Theory And Practice

Post by blueblazer »

Just a note, but my local PC recycler often gets copiers/scanners in and I have had good luck salvaging the glass platens from junked out machines.
User avatar
kempelen
Posts: 18
Joined: 07 Feb 2012, 18:19
E-book readers owned: -
Number of books owned: 1000
Contact:

Re: Platen: Theory And Practice

Post by kempelen »

Hi all,

Thanks for all instructions on this forum, my scanner is almost ready, but I have some problems, which may belong to this thread.

I tried "mitering" the acrylic, but did only half of them (too strong material). Not very well visible, but here it is:

Image

I got shadow in the middle of the book. Do you have any ideas how to fix that?

Image

Should I finish mitering or start over another way? I followed the "New standard scanner" instructions and some other ideas, but it was not clear what to do with the edges.

For most books, the shadow won't be problem. For example:

Image

But! Another question: I can see the camera shining in the middle of every page. See that image. :-(

I have one lamp, in the middle, similar lamp like in your CNC-based new scanner. It's lower than in your builds, 55cm from the bottom of the "V", but moving it upwards wouldn't fix this problem I think. Moving the camera backward fixes it, but I need to move too much out (20+ cm), until it goes out from the lamp's area. In all other scanners the camera is near to the platen and I am sure the light from the lamp hits it.

1 more, seeing this image... the upper part is lighter. Will I need two lamps pointing in a degree, instead one vertical light source? I saw this vertical solution in your "CNC based" scanner, so I thought it works OK.

Thanks in advance,
Ferenc
Post Reply