Home

Please support mflenses.com if you need any graphic related work order it from us, click on above banner to order!
SearchSearch MemberlistMemberlist RegisterRegister ProfileProfile Log in to check your private messagesLog in to check your private messages Log inLog in

Question: Ultrahigh detail picture at 15 cm
Post new topic   Reply to topic
Home   Techniques
View previous topic :: View next topic  


PostPosted: Thu Sep 23, 2010 2:24 pm    Post subject: Question: Ultrahigh detail picture at 15 cm Reply with quote
Hi all,

For a project, I need to track the edge of a sheet of paper, down to 1 micrometer accuracy. So far, not a real big problem.

But one very big problem I have, is that I can only get as close to the paper as 15 centimeters (due to a machine head, moving below the camera).

Now I wonder how I can get sufficient resolution. I'm thinking of dSLR with large zoom (expensive), or webcams with fixed magnification lenses, but I just can't find the right combination. Anyone here who can help me out?


PostPosted: Thu Sep 23, 2010 2:49 pm    Post subject: Reply with quote
Do you already have a dslr? Cheapest solution ($50-100) could be some fixed focal telephoto lens and extension tubes.


PostPosted: Thu Sep 23, 2010 5:23 pm    Post subject: Reply with quote
kansalliskala wrote:
Do you already have a dslr? Cheapest solution ($50-100) could be some fixed focal telephoto lens and extension tubes.

Nope, nothing yet. The budget is about maximum €1000.

But with a fixed focal telephoto lens, don't you get a very large working distance? Like 50-100cm? Because that's also not possible unfortunately Sad

The camera/webcam/whatever has to be positioned at about 15-30cm from the surface.


PostPosted: Thu Sep 23, 2010 5:50 pm    Post subject: Reply with quote
Welcome bloontje!

This project sounds like a machine vision application. Suggestion is go to http://www.edmundoptics.com/ and peruse the 'machine vision' section, the technical library, and contact the helpful engineers there. Most likely they offer an off-the-shelf system for your needs. Benefits include easy integration with other aspects of your application such as automation.

specific links: http://www.edmundoptics.com/applications/machine-vision-inspection/ and http://www.edmundoptics.com/technical-support/imaging/electronic-imaging-resource-guide/


PostPosted: Fri Sep 24, 2010 7:20 am    Post subject: Reply with quote
visualopsins wrote:
Welcome bloontje!
..


Thanks visualopsins, those links are really helpfull! Especially the technical help section!

I looked at their product range, but still think I have a major problem Razz
If i look at something like this: 35mm VIS-NIR Compact Fixed Focal Length Lens, the details are:
Working Distance (mm) 165 - ∞
Field of View, 2/3" Sensor (mm) 39.3
Field of View, 1/2" Sensor (mm) 28.5
Field of View, 1/3" Sensor (mm) 21.4

So when I have a 1/3" camera, i get 21.4mm of covered area. That's 21400um. So when I have like 10MP camera, it will be 4000pixels, resulting in 5um per pixel. I think it will be doable by software techniques, using the different greyscale levels in a pixel, te get an accuracy of 1um.

But the problem is, i can't find a 1/3" cameralens, with a resolution of 10MP. (and in my budget. after the lens there's 500$ left).


PostPosted: Fri Sep 24, 2010 7:45 am    Post subject: Reply with quote
bloontje wrote:
So when I have a 1/3" camera, i get 21.4mm of covered area. That's 21400um.


Embarassed .. I thought that 1 micrometer was 1/10 of a millimeter. And I have studied mechanical engineering - year 1983. I somehow got confused because the equipment called 'micrometer'.

Still, you can get very good 'macro' results with telephoto lenses and extension tubes and get about 15-30 cm distance from the subject. That all depends on the lenght of tubes and the focal lenght of the lens. And if you want to shoot later something like birds you still have the dslr and the lens. Smile


PostPosted: Fri Sep 24, 2010 4:43 pm    Post subject: Reply with quote
How about placing one lens (possibly a very small one, e.g. single element) right next to the paper itself and photographing through that lens from a distance? Would this fit into the setup?


PostPosted: Fri Sep 24, 2010 6:33 pm    Post subject: Reply with quote
The engineers at EO might be real helpful Wink Especially now you have done some homework.

There are long-working-distance microscope objectives, lwd being ~22mm

Given machine head travel path, is there room to place a tube to get the lens closer to the paper?

Can mechanics be utilized to move the lens close for measurement, and raise the lens before the machine head moves through? The measurement needs to be taken and verified okay before the machine head moves through, right? A similar approach if the paper bed moves under the machine head is to measure at smaller wd before moving the paper using precision mechanics to underneath the head.

Is the one micrometer accuracy figure smaller than is actually necessary? I don't know if paper cutting machines are as accurate. Also, coefficient of expansion by humidity of paper changes dimensions more than one micrometer, depending of course how large is the sheet; larger sheet, more dimensional change.

Only $1000 budget?! For one micrometer accuracy, I'm guessing 10x that amount might be more appropriate.

More information about the mechanical setup and purpose of the entire 'machine' may help generate other ideas. Such as is the process to become fully automated, i.e., load paper in a tray & press start? Or is this a relatively manual process where paper placement is by hand-adjusting a stage? Measure only one edge of the paper, all four edges, or the entire sheet?


PostPosted: Sat Sep 25, 2010 7:57 am    Post subject: Reply with quote
bloontje wrote:
So when I have a 1/3" camera, i get 21.4mm of covered area. That's 21400um. So when I have like 10MP camera, it will be 4000pixels, resulting in 5um per pixel. I think it will be doable by software techniques, using the different greyscale levels in a pixel, te get an accuracy of 1um.


Hmm .. someone please correct my "calculations" but do you need about 5:1 magnification then?

Something like this would then be only 2 x over your budget with something like 550D.

http://www.canon.co.uk/For_Home/Product_Finder/Cameras/EF_Lenses/Macro_Lenses/MP_E65mm_f2515x_Macro/

EDIT: and note: it IS a mf-lens Wink
EDIT EDIT: where are you located, you can rent them also
http://www.lensrentals.com/rent/canon/lenses/macro/canon-mp-e-65mm-1-5x-macro


PostPosted: Thu Sep 30, 2010 1:15 pm    Post subject: Reply with quote
visualopsins wrote:

Given machine head travel path, is there room to place a tube to get the lens closer to the paper?


Can mechanics be utilized to move the lens close for measurement, and raise the lens before the machine head moves through?

Is the one micrometer accuracy figure smaller than is actually necessary?

first of all: thanks for your help, and good thinking!
There's not place for a tube, to get the lens closer to the paper.
The mechanics thing could work maybe, haven't thought about that. I'll figure it out! Very Happy

The one micrometer accuracy is realy necessary, I want to track a sheet of paper moving through a machine, with that accuracy.

But now, I found some interesting telecentric lenses on the internet, especially this one:
http://www.couriertronics.com/docs/Sill_Datasheets/s5lpj/S5LPJ1304.pdf

It looks very awesome to me, with 18cm working distance, you get an image of about 1x1mm on a 1/3" CCD. Sounds great, doesn't it?
I don't know the price yet (think it will be very expensive, lol). But is it true that this would fit my needs perfectly?


PostPosted: Thu Sep 30, 2010 2:05 pm    Post subject: Reply with quote
bloontje wrote:
But now, I found some interesting telecentric lenses on the internet, especially this one:
http://www.couriertronics.com/docs/Sill_Datasheets/s5lpj/S5LPJ1304.pdf


aargh .. why does the register distance have to be 40 mm. Would it have been 45,5 or more you could have attached it directly to almost any srl. Smile


PostPosted: Thu Sep 30, 2010 2:08 pm    Post subject: Reply with quote
kansalliskala wrote:
bloontje wrote:
But now, I found some interesting telecentric lenses on the internet, especially this one:
http://www.couriertronics.com/docs/Sill_Datasheets/s5lpj/S5LPJ1304.pdf


aargh .. why does the register distance have to be 40 mm. Would it have been 45,5 or more you could have attached it directly to almost any srl. Smile


Hmm.. I think it's no problem to connect it to a camera like
http://www.mightexsystems.com/family_info.php?cPath=1_170_101&categories_id=101

But I'm getting more and more affraid the telecentric lens will be $5000+ :p


PostPosted: Sat Oct 02, 2010 8:57 pm    Post subject: Reply with quote
What about using a film camera? It would be much cheaper and there are fine-grained films with a higher resolving power than a digital sensor. Also you have to consider both the resolving power of the sensor/film and the magnification/reduction of image size due to the lens. If the image is magnified, the sensor/film resolution will be less critical; if the image is reduced to a third or a quarter of actual size then a sensor that resolves 5microns will actually only be recording objects of 15 to 20 microns size.

68mm of extension with a Summicron 90 lens gives about 2/3 life-size and a focus distance of about 16cm. That might work for you either with a fine-grained film or with a digital sensor with a very tight pixel pitch (what matters is surely not how many megapixels there are overall but how close together they are, so a high-end 5D Mk2 would be much worse than a cheaper 40D, for example.

Rollei Ortho 25 film http://www.digitaltruth.com/products/rollei_tech/ORTHOeng.pdf has a resolving power of more than 300 lp/mm, according to the makers, which probably gives you three or four times the power of any commercial digital sensor but means that you need an extremely good lens to avoid losing resolution there.


PostPosted: Sat Oct 02, 2010 9:32 pm    Post subject: Reply with quote
Oh - and forget the Canon 5x lens. The working distance is a couple of centimetres


PostPosted: Sun Oct 03, 2010 7:45 am    Post subject: Reply with quote
PaulC wrote:
Oh - and forget the Canon 5x lens. The working distance is a couple of centimetres


?

Quote:
Closest focusing distance: 0.243m 5X


PostPosted: Sun Oct 03, 2010 9:55 am    Post subject: Reply with quote
kansalliskala wrote:
PaulC wrote:
Oh - and forget the Canon 5x lens. The working distance is a couple of centimetres


?

Quote:
Closest focusing distance: 0.243m 5X


They don't measure it from the end of the (enormously long) lens. I think it is measured from the position of the sensor (i.e. from object to image). Trust me, I've got one.


PostPosted: Sun Oct 03, 2010 4:41 pm    Post subject: Reply with quote
yes, of course Embarassed
Post new topic   Reply to topic
Home   Techniques
Page 1 of 1