[Petri.J]

I could go somewhere.. France or Spain?
I could just relax, do nothing and enjoy my vacation..
Naeh, I'll start making an imagebased 3D scanning software!


So, to make my life easier when writing the software I thought it would be best to make stereotropic lens for my camera. There are some 3d lenses available, but they are not very good, and they are actually quite expensive. So I'll make my own.

Objective:
There is two small right angle prisms (90 degrees) in front of cameras lens that split the final image in half. After those small prisms there is a similar prism on both end that bends the light back about 90 degrees, and in result two images is projected on cameras screen that has been taken from 8 centimeters distance from each other. See right half of following picture.


Anyway the first picture is just a sketch of what that objective could look like.

On the second image left, I calculated distance of a pixel from View1. Note that I left out Y axis to make things more simple.

I could spend my time better now couldn't I?

[Sumaleth]

Cool project. I look forward to seeing how it develops.
_________________
Art Links Archive -- Artists and Tutorials

[Petri.J]

Alrighty then!
Little update here.. I ordered those prisms and tried out how they work,
and it seems like they don't. Well at least not as well as I thought they would.
Problems:
- How to do the casing?
I figured that the best way of doing the casing is with 3D printer. At first I
thought that bending it from aluminum would be the best approach, but
that would be very tricky to execute because casing has to be so accurately
crafted, since it's only like 12 centimeters wide.
- Camera's FOV
FOV causes problems, and because of it, the second prism has to be very
large. Fortunately camera has zoom so that decreases the problem a bit
but the test shows not enough. Solution to this could be to add one lens
between cameras lens and the first prism, that would decrease FOV.
Though the problem with that would be that it costs more money, and
with that in mind, It's coming clear that making this objective is
becoming less and less of a good choice.

Conclusion:
So I ended up with the most simple answer. Buy a second camera.
I will still have to make some sort of stand for those cameras,
a mechanism that allows me to take a picture simultaneously with
them and perhaps a way to zoom with both cameras.
The upside is that, with second camera, I will double the resolution.

Oh, and I thought that since I don't know C++ yet, this would be a good
opportunity learn that as well. Learning C++ could give a lot of new
opportunities to do lots of cool stuff that I have always wanted to do,
like write mental ray shaders for XSI. I've never liked the skin shader in
XSI so with C++ I could finally be able to do my own