Stereo macro with one shot

Graham Stabler · Post by **Graham Stabler** » Mon Dec 15, 2008 6:35 am

I've been taking high speed macro video for a while now and recently the people I borrow the camera from have upgraded to a Phantom V12 which has almost twice the sensor size and identical pixel size.

I wondered about trying to do some form of stereo macro, capturing both frames on one sensor.

I found this device:

http://www.loreo.com/pages/products/lor ... rocap.html

but wonder what the other options are. Preferably it would fit on the end of the Sigma macro lens I use already and maintain as much of the aperture as possible. Light is something I don't have a lot of.

Cheers,

Graham

Harold Gough · Post by **Harold Gough** » Mon Dec 15, 2008 6:52 am

Graham,

These people are the specialists:

http://www.widescreen-centre.co.uk/Cata ... ality.html

It seems useful to include this here:

http://www.crystalcanyons.net/pages/Tec ... Macro.shtm

Harold

Graham Stabler · Post by **Graham Stabler** » Mon Dec 15, 2008 8:16 am

They just sell the non-macro version of the device I posted above as far as I can tell.

Graham

Graham Stabler · Post by **Graham Stabler** » Tue Dec 16, 2008 3:23 am

This is very interesting:

http://www-graphics.stanford.edu/papers/lfcamera/

You stick a micro lens array in front of the sensor and that allows stereo to be produced digitally and it even allows extended depth of field etc. There is a tech report and if you have big "pipes" a large video.

Graham

Graham Stabler · Post by **Graham Stabler** » Tue Dec 16, 2008 5:05 am

And check this out:

http://uk.youtube.com/watch?v=8ENfPYpkHp4

They use a programmable aperture to capture the light field for a similar result. No use to me but very interesting and note their simple first attempt.

Graham

Charles Krebs · Post by **Charles Krebs** » Tue Dec 16, 2008 11:28 am

Graham,

Very interesting links! Thank for putting them up here.

Graham Stabler · Post by **Graham Stabler** » Tue Dec 16, 2008 6:56 pm

After thinking more about the first system, the "light field camera" I realised that for stereo I would just need an array of cylindrical lenses, much like are used to create lenticular holograms but with a longer focal length to account for the cover glass of the sensor. This was confirmed in another paper. For two images I need a pitch of 40um which does not seem to be an off the shelf item (Edmund have some courser ones). I'll keep looking.

For none high speed use, you can use masks instead of lens arrays as the lens arrays are just to act as pin holes anyway.:

http://www.umiacs.umd.edu/~aagrawal/sig ... amera.html

I also found mention of stereo using a "biprism", luckily one of the papers is openly available for you guys:

http://mi.eng.cam.ac.uk/~cipolla/public ... eprism.pdf

This seems very attractive if I can find a suitable prism. I might try making one from perspex as a quick reality check.

Graham

Tardigrade37 · Post by **Tardigrade37** » Wed Dec 17, 2008 7:13 pm

Graham Stabler wrote:This is very interesting:

http://www-graphics.stanford.edu/papers/lfcamera/

You stick a micro lens array in front of the sensor and that allows stereo to be produced digitally and it even allows extended depth of field etc. There is a tech report and if you have big "pipes" a large video.

Graham

I attended a lecture by Marc Levoy on this topic this past spring. Some people are working to use microlens arrays in very novel ways to measure multiple parameters simultaneously. See the recent publication by Rudolf Oldenbourg:
Polarized light field microscopy: an analytical method using a microlens array to simultaneously capture both conoscopic and orthoscopic views of birefringent objects

Journal of Microscopy, Volume 231, Number 3, September 2008 , pp. 419-432(14)

augusthouse · Post by **augusthouse** » Wed Dec 17, 2008 7:45 pm

There is also an interesting video on YouTube regarding the Stanford light field camera.

http://www.youtube.com/watch?v=9H7yx31yslM

Craig

Graham Stabler · Post by **Graham Stabler** » Thu Dec 18, 2008 3:11 am

Craig, thanks that's a bit easier on peoples connections than the video on their website.

Tardigrade37, thanks for the reference!

For my stereo I am going to try a bi-prism in front of the lens, I'm having trouble finding one of these so I'm going to use two Fresnel prisms back to back. As has been recently reported (Optics Express, Vol. 16, Issue 20, pp. 15495-15505), a terrible paper to read but a good idea. These are very similar to Fresnel lenses (those flat but rough lenses used for reversing cars and as book magnifiers) except is works as a beam diverting prism. I just need to work out what angle I need to get.

Cheers,

Graham

Graham Stabler · Post by **Graham Stabler** » Thu Dec 18, 2008 3:13 am

One other thing I was going to say is that if anyone fancied trying a programmable aperture they could make the mask as a large disk with the apertures around the circumference. This could be laser cut very cheaply (in fact I have a laser cutter so would be happy to help any experimenter) and would be easy to set up and use..

Graham

Graham Stabler · Post by **Graham Stabler** » Thu Dec 18, 2008 6:53 pm

I have found a source of in expensive Fresnel prisms to recreate the biprism type stereo.

http://www.west-op.com/pressonprism.html

These sticky back prisms are made by 3M to correct for some eye problems such as double vision. They come in a variety of strengths too so I hope to work out which one I should use and order a pair. Then I cut them in half and stick one half from each pair onto a filter facing oposite ways and that should be that, at least I hope.

For choosing:

"One prism diopter is defined as the prism power needed to deviate a ray of light 1 centimeter at a distance of one meter"

Graham