Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

New Camera Inspired By Insect Eyes

samzenpus posted about a year and a half ago | from the looking-at-the-world-through-fly's-eyes dept.

Science 35

sciencehabit writes "An insect's compound eye is an engineering marvel: high resolution, wide field of view, and incredible sensitivity to motion, all in a compact package. Now, a new digital camera provides the best-ever imitation of a bug's vision, using new optical materials and techniques. This technology could someday give patrolling surveillance drones the same exquisite vision as a dragonfly on the hunt."

cancel ×

35 comments

Sorry! There are no comments related to the filter you selected.

This is wonderful (0)

Anonymous Coward | about a year and a half ago | (#43608187)

How did we all get so lucky to live in an age of technological marvels like this.

pinhole camera. (4, Interesting)

Anonymous Coward | about a year and a half ago | (#43608193)

If you make each sensor small enough with the appropriate overlay mask - you get a pinhole camera with an infinite depth of view.

The advantage an array of such cameras is the ability to integrate thousands of small images to create a 3D result.

Re:pinhole camera. (0)

Anonymous Coward | about a year and a half ago | (#43608909)

Don't forget smell-o-vision too.

Bent electronics - a first? (3, Interesting)

G3ckoG33k (647276) | about a year and a half ago | (#43608217)

From another article on the same topic

http://www.colorado.edu/news/releases/2013/05/01/insect-eye-inspired-camera-captures-wide-field-view-no-distortion-according [colorado.edu]

"“The most important and most revolutionizing part of this camera is to bend electronics onto a curved surface,” said Jianliang Xiao, assistant professor of mechanical engineering at CU-Boulder and co-lead author of the study."

So, electronics have not been bent like this before, whether for optronics or otherwise? Maybe it is too obvious, in hindsight.

Re:Bent electronics - a first? (2)

QuasiSteve (2042606) | about a year and a half ago | (#43608569)

Bending electronics is very simple, for broad definitins of bending and electronics.

See that flex cable inside your laptop? It bends, and technically it's electronics.

Okay, there's no components on it. So let's put some slightly more rigid bids on there, and populate those with components. It's still 'bendable' even if the parts where the components are do not bend.

The components themselves, of course, remain highly inflexible, but the smaller you can make them, the less of an issue this is for the overall flexibility of the system.

In the case of the 'camera' in the article, they're really, really, really tiny. But they also don't really have to bend - they're still on a rigid surface, just not a flat one (though at the scale in question, that's akin to arguing the ground you're walking on isn't flat because the entire Earth isn't.)

What they've created here is kind of a lytro/fisheye hybrid. Neat, but for general photography it would be much more interesting, short-term, if the sensor in the camera body could be curved, so that the lens and software don't have to try and correct for the flat surface quite so much, and make it affordable. Think 'curved film plane', but on both axes. There's been a flurry of patents on curved sensor designs in the past few years, so give it another decade or so?

Re:Bent electronics - a first? (1)

Skapare (16644) | about a year ago | (#43613307)

Take apart a modern high-end DSLR camera and you see much well bent electronics (no sensors on it, though).

Re:Bent electronics - a first? (2)

bkr1_2k (237627) | about a year and a half ago | (#43609161)

Flex circuits have been around for a while. That's not particularly revolutionary. We've been using them for about 5 years. They're typically fairly simple circuits, but still this isn't the "cutting edge" part of this research. Getting the components small enough to still be flexible on the 1 cm circuit may be the part that he's referring to, though, which is less trivial.

A little bit of Matrix for your Slashdot (1)

CrimsonKnight13 (1388125) | about a year and a half ago | (#43608261)

This technology definitely has me thinking about the sentinels from the Matrix. http://matrix.wikia.com/wiki/Sentinel [wikia.com]

Re:A little bit of Matrix for your Slashdot (0)

Anonymous Coward | about a year and a half ago | (#43609523)

Not funny or smart, please try again.

new camera! NO PICTURES (0)

Anonymous Coward | about a year and a half ago | (#43608283)

It's just great how they have this camera and can't provide any pictures taken by it. Very pro stuff, nothing says proof of concept like an inoperable device.

Great!! (0)

Anonymous Coward | about a year and a half ago | (#43608357)

I like pie.. I know it's irrelevant.

Imagine a Beowulf-cluster of these! (1)

Gaygirlie (1657131) | about a year and a half ago | (#43608361)

...oh, wait...

Since when are compound eyes high resolution? (2)

AC-x (735297) | about a year and a half ago | (#43608369)

Since when have compound eyes been known for being high resolution? A dragon fly and its 30,00 lenses only corresponds to a total resolution of around 200 x 150.

Compound eyes have many advantages for miniaturisation, field of view and sensitivity to movement, but there is no way you could claim they were high resolution.

Re:Since when are compound eyes high resolution? (3, Informative)

Immerman (2627577) | about a year ago | (#43610177)

What you're describing is a typical apposition eye, where each lens focuses light onto a single photoreceptor. That's actually the simplest form of compound eye, found in arthropods, annelids, and some bivalves, and it may still deliver higher resolution than the number of facets if operated as a phased array. A slightly more complicated version of the same design, the schizochroal compound eye, actually uses multiple photoreceptors per lens with the resulting hundreds or thousands of low-resolution images being (presumably) composited by the brain. (and incidentally - neuron-for-neuron insects have the most complicated brains on the planet, with each neuron making tens to thousands of times as many connections as those in a human brain)

Skilled fliers and prey-catching insects such as dragonflies typically have among the most sophisticated compound eye designs which deliver quite high resolution - if you've ever been used as a hunting perch by dragon/damselflies you can witness this - they'll be comfortably perched and then dart out to grab some tiny flier that was barely visible to your eye, even if your head was considerably closer to it than the dragonfly's perch was.

Re:Since when are compound eyes high resolution? (1)

AC-x (735297) | about a year ago | (#43610489)

Dragon flies have very good vision for what the use it for, with a higher detail / magnified area in the centre of their vision similar to the fovea of the vertebrate eye, but it's still not high absolute resolution and is well below all but the most basic conventional digital cameras' sensors.

Re:Since when are compound eyes high resolution? (2, Interesting)

Anonymous Coward | about a year ago | (#43611877)

Video of what Immerman is talking about. http://www.youtube.com/watch?v=dOyc98tV5kA

See Nova: Rise oh the Drones, January 23 2013 (1)

crackspackle (759472) | about a year and a half ago | (#43608371)

This tech was discussed there. It. was created by a guy in his garage using harvested mobile phone cameras and is already being tested in drones by Homeland Security. Per the show, the actual level of detail is classified but they did show an example where the were able to monitor dozens of city blocks at the same time grid style and then choose any point on the grid and zoom in hi-def and see full detail of people walking on the ground, all using a single lens array.

Is it real or just a concept? (0)

Lumpy (12016) | about a year and a half ago | (#43608403)

Honestly, I really doubt this kind of stuff where you can not find ANY photos of the claimed device anywhere online.

It's the same as claiming you invented cold fusion but not allowing anyone to look inside the box and to ignore the sound of a honda generator running inside the box.

Re:Is it real or just a concept? (1)

AlphaWolf_HK (692722) | about a year and a half ago | (#43608939)

Yeah if they're looking to inspire investors, this is a horrible first step.

Re:Is it real or just a concept? (1)

CanHasDIY (1672858) | about a year ago | (#43610463)

Yeah if they're looking to inspire non-government investors, this is a horrible first step.

FTFY.

DoD has a nasty habit of playing fast and loose with the purse strings.

Re:Is it real or just a concept? (1)

jwdb (526327) | about a year and a half ago | (#43608987)

Yes, the links in the summary are poorly placed, but try clicking on them anyway. You'll quickly get the Nature article with renderings, photos and tests of the device.

Definitely not a black box, and the article isn't even that difficult.

Re:Is it real or just a concept? (1)

zAPPzAPP (1207370) | about a year ago | (#43611251)

Aside from that, it is a camera.
Yet there are no pictures said camera took anywhere in the article.

When writing about an image taking device, the first thing right below the headline should be an image taken by said device.
You know, like in the old saying about pictures vs. lots of words.

lightfield cameras may work in similar way (2)

peter303 (12292) | about a year and a half ago | (#43609231)

They began as an wallisze array of video cameras a decade ago. But people learned you can get similar results with an array of special lenses on to a single large camera with a lot of computer postprocessing. The array of lenses looks like an insect eye.

Re:lightfield cameras may work in similar way (1)

Anonymous Coward | about a year ago | (#43609673)

A lightfield camera isn't quite the same, though you could think of a compound eye as a lightfield imager whose focal plane has been warped onto a sphere and whose elements are each 1x1 pixels.

Re:lightfield cameras may work in similar way (3, Interesting)

ka9dgx (72702) | about a year ago | (#43609805)

Yes, this is just like a plenoptic (light field) camera. If you want to experiment yourself, all you need is non-moving subject material, a digital camera, and time. Take photos from slightly shifted viewing positions of a subject. Then use Hugin or Photoshop to align them on a chosen subject (or focal plane). Average all the frames together, and you'll have a synthetic focus image of your subject.

With some care and effort, you can even supersample the pool of images and get super-resolution output, where the result is more pixels than any source image (but far less than the sum of all the images).

I've been doing experimentation along these lines for a few years, and here are the resulting photos [flickr.com] of scenes from the Chicago area. I was inspired by the work of Marc Levoy, and his Stanford Multi-Camera array. [stanford.edu]

Re:lightfield cameras may work in similar way (2)

peter303 (12292) | about a year ago | (#43609997)

I've been observing developments at SIGGRAPH for over a decade. Fanstastic stuff! It used to cost over $50K to build a wall (16) of old-style video cameras with computers to process the images.

All this from asking the "what if" question: "what could you see if you had more than two eyes?" i.e. a wall or hypercube of arbitrary close eyes. Some very interesting answers in the SIGGRAPH papers. Some answers not at all expected.

Just give it time (1)

Anonymous Coward | about a year and a half ago | (#43609543)

There's really no need for research and debate on this. It's been very clearly "proven" that evolution will allow things to gain information and improve all on their own. So, just let the camera sit for a few million years, and the problem will solve itself. After all, that's the "scientific" approach...

Re:Just give it time (0)

Anonymous Coward | about a year ago | (#43609691)

You obviously know very, very little about the way evolution by natural selection works.

Re:Just give it time (0)

Anonymous Coward | about a year ago | (#43611761)

You obviously have a broken humor circuit, and are liked by no one.

Re:Just give it time (1)

Entropius (188861) | about a year ago | (#43610227)

Natural selection requires a thing that makes imperfect copies of itself. A camera sitting on the shelf doesn't do that.

Re:Just give it time (0)

Anonymous Coward | about a year ago | (#43611311)

not the same camera, certainly...but that camera will somehow make other cameras similar but better than itself...with no need for any guidance whatsoever... pre-existent matter, pre-existent energy, pre-existent physical laws...the normal stuff of "provable science", not that crazy religious stuff that relies on things you can't prove empirically.

Mantis shrimp eyes next, please (1)

shilly (142940) | about a year ago | (#43610939)

I wanna know when they're going to create artificial mantis shrimp eyes. That I can put on and use. That would be cool, given that they are the most impressive eyes on the planet. I remember reading Fragment by Warren Fahy and being blown away by what they are capable of.

http://en.wikipedia.org/wiki/Mantis_shrimp#Eyes [wikipedia.org]

Ah, yes (1)

ThatsNotPudding (1045640) | about a year ago | (#43611617)

Three-Dimensional Freedom (TM)

No, wait: the exact opposite of that.

Perception difference, photos irrelevant (1)

Yoik (955095) | about a year ago | (#43612651)

Compound eyes are different in a deep perceptual sense than mammal eyes, and pictures capture the content of mammal perception. A compound eye's perception would be great for a robot to use for navigation, as it provides info for a 3d model of the environment with rapid identification of any moving features. Mammal eyes are better at resolving details of features. The trade offs can be reconciled with mammal eye movement and processing.

One problem humans have, is easy understanding of what a compound eye "sees" and how to process it. We have good intuition about how to capture the images and process sequences of them. Not so with a compound eye that isn't intended to capture an image. It will take a while to develop that understanding.

Check for New Comments
Slashdot Login

Need an Account?

Forgot your password?