Camera for stereo vision using mirror(s)

Forums › Digital Imaging › Camera for stereo vision using mirror(s)

noteb

5 posts

Wannabe Geek

#154463 28-Oct-2014 22:38

Hi all,

For a stereo vision project I am looking to by a high quality mirror about 4" size.

Could anyone please point me to the right direction.

So far I've been able to find only this one http://www.astronomy.co.nz/shop/item.aspx/secondary-mirror-41mm/138/ , but neither shape, nor thickness are suitable.
The shape is better to be pyramidal (or rectangular) and the thickness - as thin as possible as the weight matter.

Thanks.

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1164063 29-Oct-2014 03:44

You wanting a mirror? Or a prism?

Quic Broadband

Move to New Zealand's best fibre broadband service (affiliate link). Note that to use Quic Broadband you must be comfortable with configuring your own router.

noteb

5 posts

Wannabe Geek

#1164085 29-Oct-2014 07:12

I think a mirror would be better for several reasons. Firstly, the weight of a prism of needed size. Secondly, one prism won't suffice as it would split the the ray by frequences. And third, a good quality prism of that size would be more expensive than a mirror, I suspect.

Please correct me if I'm wrong.

Fred99

13684 posts

Uber Geek

#1164113 29-Oct-2014 08:42

Google "front surface mirror" or "first surface mirror" (mirrors "silvered" on the top surface).
Prisms are a normal way to do this - that's how prismatic binoculars work.

noteb

5 posts

Wannabe Geek

#1164555 29-Oct-2014 18:40

Thank you all for replies.

Porro prisms in binoculars have light entering and leaving normally and reflected inside. That is why it is not dispersive. However, in order make a stereo system to work with prisms (and not have huge prisms) there will be lenses required and more than one prism. So it keeps growing in size and weight. Little weight is essential in the case of my project that is why I want to go with just one mirror. Although a good part of the observable world is lost in this case it is still enough to get a depth map.

Where to buy a first surface mirror in New Zealand? I haven't found any suppliers.

Thanks.

eracode

Smpl Mnmlst
8766 posts

Uber Geek

ID Verified

Trusted

Subscriber

#1164753 30-Oct-2014 02:36

Stereo vision? = 3D? Would be very interested to know what your project is, if it's not a secret.

Sometimes I just sit and think. Other times I just sit.

noteb

5 posts

Wannabe Geek

#1164756 30-Oct-2014 02:57

It's not a secret but I would be very interested to know where to get a first surface mirror in New Zealand :)

The project is a UAV for interiour use. There are number of challenges to this including orientation in a building. For that after a lot of digging I decided to go with stereo vision to get a depth map with much more possibilities in the future. Other types of sensors are not good as I plan to make multiple such vehicles to work together and they will simply mess with each other's sensors. Those sensors include active depth map sensors (e.g. kinect) and sonar sensors (don't even provide enough information).

Fred99

13684 posts

Uber Geek

#1165021 30-Oct-2014 14:09

I can't help but think that unless there's a good reason not to do so (like cost / required quality of the camera/imaging device) then using two separate inexpensive cams may be much less trouble to set up and keep aligned than setting up frames etc to hold mirrors in place.

noteb

5 posts

Wannabe Geek

#1165084 30-Oct-2014 15:18

There is a good reason for not using two inexpensive cameras - synchronisation. If the cameras are not synchronised in hardware then one camera will see slightly different scene that the other camera making it impossible aquire a correct depth map. Synchronised solutions cost too much.
Synchronisation must be on frame capture and not just clock (crystal oscillator) synchronisation.

Thanks anyway, now I know what type of a mirror I need. I did try using regular mirror - it's a disaster - a bit blury and slightly different spectrum.