h a l f b a k e r yRenovating the wheel
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Instead of a single lens, the camera uses 100 small lenses and CCDs packed into an array. It uses cheap mass produced plastic lenses (the type used in disposable cameras) and cheap, low spec CCDs
When you take a picture, some clever piece of software combines the 100 low quality images to produce
a single high quality image. Defects in the lenses and CCDs are subtracted out of the underlying image by comparison of the large number of pictures.
The advantages:
* Can potentially take very high quality images, but is cheaper to manufacture than a professional camera. a) The relationship between CCD price and CCD size is non-linear b) Does not need a high spec lens.
*The light-capturing cross-section is what fundamentally limits image quality for short exposures. This camera has a high cross-section but does not require large and heavy lenses.
*Flatter than a comparable camera with a single lens.
Scientific paper on same
http://www.suss-mic.../MOC-04_Duparre.pdf
[DrCurry, Jul 18 2006]
Shoot first, focus later
http://www.newscien...d=online-news_rss20
Using 90,000 lenses [fridge duck, Jul 18 2006]
Ah, just how to make a compound eye
http://underdone.mu...com/journal/item/63
...not a camera. [DrCurry, Jul 18 2006]
Light field
http://en.wikipedia.org/wiki/Light_field
Different technology achieving the same end as this [mindplay, Jun 22 2011]
Lytro Cameras
http://www.lytro.com/cameras
First (?) commercial application of light field sensors [mindplay, Jun 22 2011]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
There was an article in the NY Times not too long ago about using the compound eye approach for cameras. Not sure I'll be able to find it. |
|
| |
Not only does this allow the benefits that you stated, but it also allows depth of field to be adjusted as a post-process by simply choosing which feature you wish to be in focus, and aligning the multiple images to match. Also, since each image is captured from a slightly different point, it can be used to generate a depth map (3-d imaging). |
|
| |
If it weren't for the fact that this technique was already in the oven with the burners lit, I'd bun it. |
|
| |
I don't know what kind of quality will be possible if the image is going to be made by low quality optics - but on the subject of post-exposure focussing, have a looksee at the link. |
|
| |
//I don't know what kind of quality will be possible if the image is going to be made by low quality optics// |
|
| |
If each lens is slightly different, in theory the defects of the lenses should cancel out. I am assuming that low quality lenses are capable of producing high quality images within a narrow range of parameters. |
|
| |
I could imagine, for example, using only one colour channel per lens, and placing red/green/blue filters in front of a third of the lenses respectively. That way, chromatic abberration is eliminated and there is no need to use exotic materials. |
|
| |
I just had pretty much the same idea today, while
studying the camera on my cell phone, and I
thought I'd take a look here to see if somebody
had already had the same idea. |
|
| |
In my opinion, you would not need that many
lenses - I was thinking 2x2 (4) or perhaps 3x3 (9)
lenses in total, and these would have to be placed
very closely to each other... using pinhole lenses
and tiny CCDs, it should be possible to place them
as closely as 1/2 or maybe 1/3 inch apart. |
|
| |
Using wavelets or fourier-transforms to
extrapolate the image data from 9 images, even a
4-lens configuration should be enough to get fairly
high precision, but with 9 lenses, I would say, the
odds that 4 or more lenses have the same error for
any given pixel is infinitesimal. |
|
| |
The main reasons I thought of this idea myself
were not price or portability, but rather: |
|
| |
1. Pinhole lenses have an extremely wide focus
range - which of course is going to be somewhat
reduced by placing them in an array, but this can
be partially corrected in software. |
|
| |
2. I thought it would be useful and interesting to
be able to take high-quality photos where
subjects in both the foreground and background
are in focus - this is not easy with traditional
optics, at least not without using very slow
shutter speeds. Using many lenses will capture
many times more light, thus higher precision at
higher shutter speeds. |
|
| |
Both of these qualities would make a camera of
this type great for video recordings too - due to
the high shutter speeds, you would be able to
capture video with very little motion blur. |
|
| |
One last thought: perhaps in a 4x4 lens array, if say
6 of the 16 lenses were infra-red (and the camera
was fitted with powerful infrared diodes), this
could enhance the capability to shoot still images
or video at night... Night photography can be
tricky - but with a camera like this, you don't need
to worry about adjusting the focus. |
|
| |
This is an interesting idea for sure! |
|
| |
Any idea if this is currently being researched by
any camera manufacturers? |
|
| |
I'm pretty sure I've seen a 5x5 compound lens camera under development for mobile phones.
I'll see if I can find a link. |
|
| |
Hexagonal lenses would give a better packing fraction, and any
distortion could be removed by postprocessing the image
data. |
|
| |