Quantum Ghost Imaging

This is cool:

Ghost imaging is a technique that allows a high-resolution camera to produce an image of an object that the camera itself cannot see. It uses two sensors: one that looks at a light source and another that looks at the object. These sensors point in different directions. For example, the camera can face the sun and the light meter can face an object.

That object might be a soldier, a tank or an airplane, Ron Meyers, a laboratory quantum physicist explained during an Oct. 28 interview on the Pentagon Channel podcast "Armed with Science: Research and Applications for the Modern Military."

Once this is done, a computer program compares and combines the patterns received from the object and the light. This creates a "ghost image," a black-and-white or color picture of the object being photographed. The earliest ghost images were silhouettes, but current ones depict the objects more realistically.


Using virtually any light source -- from a fluorescent bulb, lasers, or even the sun -- quantum ghost imaging gives a clearer picture of objects by eliminating conditions such as clouds, fog and smoke beyond the ability of conventional imaging.

EDITED TO ADD (12/12): A better explanation of the effect, and a detailed paper.

Posted on November 18, 2009 at 6:22 AM • 30 Comments


kevinmNovember 18, 2009 7:04 AM

That URL should carry a "best viewed with NoScript" health warning. When I viewed it there were three flashing adverts and a sound track saying "congratulations you won". When I see that it greatly reduces my trust in the site.

Paul CrowleyNovember 18, 2009 7:48 AM

The "more detailed link" above makes it clear that this is *not* a quantum technique; I would have been astonished if it had been genuinely quantum, given your description of the equipment required.

RandallNovember 18, 2009 8:07 AM

Well, the more detailed link shows that some physicists think that the effect is not quantum. If the described mechanism is working as the researchers believe it is, though, it is quantum.

MoNovember 18, 2009 8:21 AM

Cool. Like in Blade Runner where they have the cameras where you can zoom in and look around corners.

DerobNovember 18, 2009 8:41 AM

Somehow this sounds to me like optical snake oil.
One of the things which makes me wonder is the fact that the 'paired' photons need to reach both detectors at the same time. They are however traveling via different paths with different lengths. As the speed of light may be quick but certainly not infinite, they would typically not hit at the same time.
An other issue that is a bit under explained is the practical application if you would still need a sensor pointing at the object. It is not the point that this could be a CCD camera, since a light meter does not have to focus (it is single pixel anyway), but the fact that there needs to be a device in the first place.
By reading the detailed paper (thanks kevinm) it becomes clear that the image actually being calculated by subtracting the knowledge about the light source. It could indeed prove a way to see through smoke or fog (which makes focusing useless), and this certainly sounds like a useful application if it works outside of the laboratory.

jgrecoNovember 18, 2009 8:53 AM

I don't know if this observation is worth anything or not, but their setup reminds me heavily of how holograph photography works.

Shai MachnesNovember 18, 2009 9:36 AM

This is not a quantum phenomena at all.

Given a light source with a known non-homogeneity in space and time, you can deduce the shape of the object from the total amount of illumination reflected from it. That actually makes sense.

The "Quantum" in the name stuck, even though the original explanation is wrong (as proved in the experiment linked to below).


monopole November 18, 2009 9:44 AM

Not quantum, basically an "extreme" form of compressive single pixel dual photography


very neat stuff with implications for imaging through scattering media imaging of objects not in field of view etc.

This is all a subset of computational imagery which has immense security applications

Imagine camera arrays that can image through scattering media, shift focus and dynamic range, transition into high speed cameras or high resolution cameras simply by changing code.

kashmarekNovember 18, 2009 10:54 AM

Why does the phrase "smoke and mirrors" come to mind? Will such illuminated pictures stand up in court?

MarkHNovember 18, 2009 11:14 AM

Instead of going through all this expense and trouble to end up with an image of a silhouette of the object you didn't take a photo of, why not just take a photo of the object?

I suppose this technology is interesting, but until and unless it can produce an image of the object that's *better* than whatever results from taking an actual photo under whatever sub-optimal conditions prompted this exercise in the first place...meh.

RHNovember 18, 2009 11:48 AM

okay, the detailed one is dense, but the "better" explaination begs a question. If you need entangled photons, don't you need a beamsplitter? Its hard to beamsplit the sun!

On the other hand, it could be useful to minimize losses due to distance. Currently reflected light is affected by R^4 losses (R-squared towards the surface, then R-squared coming back, assuming uniform scattering off of the object). I don't see why you couldn't build one of these splitters on an aircraft, bounce one photon back and forth while the other one heads out. You'd need a lot less brightness to illuminate the scene because you'd have a fraction of the losses.

I think it could also be used to build topo-maps by varying the length of the reference beam.

ShaneNovember 18, 2009 1:36 PM

@Ward S. Denker

And the majority of humanity's technological breakthroughs are different how...?

RHNovember 18, 2009 1:50 PM

Correct my logic if I'm wrong, but there's R^2 losses as you emit light from a plane. When it hits, it then radiates in all directions, generating R^2 losses back. The result should be 1/R^2 * 1/R^2 losses which is 1/R^4.

I think you have the correct formula for the case where you have a mirrored surface reflecting the light directly back - the case I am referring to assumes just scattering in all directions. Let me know if I missed something (I'm sort of stealing the R^4 from radar laws, I think they apply to light)

JonSNovember 18, 2009 2:30 PM

@MarkH "Instead of going through all this expense and trouble to end up with an image of a silhouette of the object you didn't take a photo of, why not just take a photo of the object?"

This is potentially another way of seeing in the dark. Point the camera at the moon, or a streetlight, or a star, and point the sensor where you think the enemy is. Et voila - there he is, naked to observation like a new-born babe.

averrosNovember 18, 2009 4:27 PM

@Shane: "And the majority of humanity's technological breakthroughs are different how...?"

They were made for entirely peaceful purposes, as obvious to any student of history of technology - and those with military applications were typically "weaponized" much later. A real military (as opposed to the armchair warriors in US military command - which lost last 30 or so wars to seriously weaker opponents) is rather conservative - no soldier or general wants to trust his life to new gadgets with unknown tactical properties.

Despite what the military-industrial propaganda tries to sell, the returns from military research are negative. The claims that somehow war and war making are beneficial economically are nothing more than a good old Broken Window fallacy. (Look it up).

Filias CupioNovember 18, 2009 4:53 PM

I agree with RH: R^-4 applies to using a camera with a flash in darkness on a diffuse-reflection target. (2R)^-2 would apply for light returned from the flash by a mirror. Using uniform ambient light (normal daytime photography) you get R^-2 (via the extremely good approximation that all objects in sight are effectively the same distance from the sun.)

In the case of radar, you get to use the same antenna for transmission and detection, so you get your antenna gain twice. This doesn't normally apply to photography.

bwpNovember 19, 2009 7:43 AM

@averros - They were made for entirely peaceful purposes, as obvious to any student of history of technology - and those with military applications were typically "weaponized" much later.

That's not really true at all. The research might have been "weaponized" later but the funding for it was, for the most part, always being paid for by the military for the express purpose of helping with war. A lot of technology never sees civilian hands until it's been surpassed by newer technology or dumbed down enough that a civilian force couldn't match the military.

BrunzjinffeNovember 19, 2009 10:27 AM

@Filias Cupio,

> In the case of radar, you get to use the same antenna for transmission and detection, so you get your antenna gain twice.

Only for some kinds of radar; not if you're trying to overcome stealth. (Security Engineering 2, p578)

averrosNovember 20, 2009 4:05 AM

bwp - that's what the military-industrial propaganda tells you. This is a relatively recent notion, post-WWII.

Of course, they forget to tell you what research DIDN'T happen because of resources consumed by military research. Yes, time of smart people and money taken from their potential backers by force to finance the military research. Even with that, the closer examination of any purported military-made wonder technology usually shows that it'd happen anyway, without any heads in funny caps around.

Please, look what the "Broken Window" fallacy means - you're committing it by uncritically repeating propagandist claims.

neillNovember 22, 2009 2:41 AM

NASA (and other astronomy research) uses those effects for decades now
e.g. you can't see a distant planet but its effects on other objects closeby

there was some other research on how to detect stealth planes by detecting them blocking satellite signals

Clive RobinsonNovember 22, 2009 6:17 PM

@ neill,

"there was some other research on how to detect stealth planes by detecting them blocking satellite signals"

It is now no great secret and is fairly obvious if you thing about it.

To be fully stealthy you need to do two things,

1, Stop reflections.
2, Allow transmission.

That is in a black room with just a single candle lighting it you can hide by wearing black (absorbant) material with a mat surface (no reflection).

However walk between the candle and an observer then you will be seen by your shadow (transmission loss).

Then there is the question of if your black suit is the same black as the backdrop across the entire EM spectrum.

The current state of practical stealth technology is "absorb" and minimise "reflection" back along the EM source path.

The UK's manufacturer of the rapier missile system developed a number of working systems to detect the US stealth fighter/bomber and gave a working demonstrated at an international trade show.

There are infact several currently known ways that stealth technology can be defeated.

The mechanical design of a stealth aircraft is such that any reflections should be minimal back towards the EM source (no curves, or 90 degree angles and plane surfaces not normal to the likely direction of an EM source). However "offset" or space diversity EM systems with multiple radiators and receivers will pick up reflections off of plane surfaces that are not reflected normal to the plane surface (ie 90 degrees to the plane thus 180 degree reflection back to the radiators).

To limit this the aitcraft surfaces are coated with EM absorbing materials. However they are by no means perfect in the same way matt black paint is not perfect.

Firstly the abosrbtion is not uniform across the full EM spectrum. And secondly an absorber will actually reflect at certain angles, you can see this with a flat board painted matt black as you tip it through the light at angles close to it's plane it appears shiney.

Thus even in the part of the EM spectrum where the absorber works it will still have angles at which it reflects. Obviously by grading the surface (as in the glass of optical cables) you can minimise this reflection. But doing this is a little like having air bubbles under freshly put up wallpaper where you try to push it down it makes it worse somewhere else (worse a graded surface can act like one of those edge lit emergancy exit signs that give illuminated words in the middle of a pane of glass.

Importantly any reflecting surface changes the phase of a coherant EM signal near it's surface, this produces peaks and nulls in the field where the non incident EM radiation and reflection co exist. If the reflecting object moves with respect to the EM source then the phase will change, this causes incidental FM modulation of the field and then there is also the doppler effect as well.

Also the peaks and troughs move making the signal AM modulated (think moon light reflecting off of rippling water). Thus all three modulation components are related to each other which alows cohearent demodulation of the reflected signal.

Further the combined effect is to widen the bandwidth of the combined EM signal at the receiver. Now by using some fairly well known techniques you can "null out" the direct path EM signal from the source and leave the modulated reflected signal.

Combined, space diversity systems and advanced signal processing techniques can significantly reduce the level of stealth the curently implemented stealth technology produces.

One of the curious things that happens at the edges of an object when it is illuminated from behind or where the EM source is not directly visable to the receiver, is what is sometimes called the "edge effect" or "skirt effect" which effectivly bends the EM waves and enables the receiver to pick up the EM source. Usually the sharper the edge the more pronounced the effect, however "sharp" is relative to the EM wavelength.

There are other issues to do with turbulance coming off of an aircrafts wings and other surfaces. Under certain conditions this turbulance can be detected by the effects it has on parts of the EM spectrum (in a vaguly similar way to the turbulance seen under water from a propeller or impeller, or in clouds with certain type of weather radar).

This is all before transmission loss effects (hand infront of your face / shadow). And this is the real problem with current practical stealth systems they are the equivelant of the "black Ninja suits" not a mythical "cloak of invisability".

To do "invisability" you have to solve the transmission loss problem by being either losslessly transparent across the whole EM spectrum or cause it to wrap around you coherantly in all directions with zero loss and distortion. Neither is much of a proposition currently (which is why stealth aircraft fly at night ;)

Which means any object that gets between an EM source and a receiver is going to be detected by a loss in signal at the receiver. But also it will throw a shadow onto a back drop which can be detected. So in theory you could put suitable EM sources and receivers on a geo stationary satellite system with a tight focus covering an area of the earths surface. Two or more satellites monitoring the earths surface from sufficient angle would observe the shadows made by all aircraft. Is it practical as a system well... Let's just say we have the knowledge and technology to do it, if we can get the required CPU power to make it practical ;)

JakeNovember 23, 2009 4:18 AM

hundreds of words, thousands, tens of thousands ... my kingdom for a pictorial demonstration?

HesoNovember 23, 2009 4:20 AM

@averros: You're comitting a fallacy of your own. The technological advances made by the military are prompted by the urgency of defeating the enemy or surviving as a nation at times of war, or by hostility and paranoia in times of peace. These factors are not purely monetary.

Yes, the money spent during war on research could be spent just as well during peace times. But the motivation is not there, nor is the universal focus, the urgency.

Besides, I don't know what history you are reading. It is a matter of record that many technological achievements were produced during war. It should have signalled you that your use of the broken window fallacy is wrong.

Take the computers, whose development was hastened by the struggle to defeat the Enigma machine during WWII; the Internet, which was created by the US military as a strategic project (a data network without single points of failure); the advances in rocket science by the Germans during WWII were cannibalized later by both Americans and the Russians, enabling them to develop their space programs. And so on.

Clive RobinsonNovember 23, 2009 5:11 AM

@ Heso,

"Take the computers, whose development was hastened by the struggle to defeat the Enigma machine during WWII; the Internet, which was created by the US military as a strategic project"

For the historical record the DOD project from which TCP/IP originated had little or nothing to do with WWII. I'm not sure if you ment to imply it did but the context of your post makes it read that way.

With regards computers the story is a little more complicated.

The military have always had a need for computing devices (calculators) for weapons aiming and delivery systems.

There where many programable calculators (tabulating machines) available prior to WWII based on the requirments for census office work (look up the history of Holerith "strings" and tabulation in general).

What went on at Bletchly Park with the development of the bombs etc is a little complicated.

There is evidence to show that prior to going to Bletchly Turing had put all the pieces together for a general purpose computing engine however the reliability of valves and their power requirment precluded him going down that path and his experiments with electro mechanical devices did not look that promising.

What did happen at Bletchly was the work of Tommy Flowers and others from Dollic Hill in showing that used properly valves where quite reliable, and the first machine was an electronic comparitor to search through "Fish" traffic looking for the likley keying point (the "fish" system unlike Enigma was effectivly a stream cipher).

It was the success of this that made Turing and others such as Gorden Welchman realise that the use of thermionic valves was practical.

So in a way the only effect WWII had on computer development was letting it move down a path that did not end in a "mechanical" dead end as had Babbages "mill" work a hundred or more years before.

The other thing that WWII did do was act as a wakeup call to the Military and Politicos about Science.

WWI had been the first war where science (not manufacturig which was the US war of independance) had had a major part to play. However the military mind was slow on the uptake and the politicos very pointedly did not want to know (see about Churchill and building warships to run off of oil not coal).

WWII was a wake up call because it proved beyond boubt that major warefare was now won with science not battle field set pieces and castle mind set defences.

Oddly though the various Afgan conflicts have shown that gurella warfare and determination beats technology hands down every time. Which is possibly why the US has it's air conflict and action from a distance theories, which pointedly do not work when the threat is in your own homeland (terrorism / freedom fighting) for which the military have no response that will work.

Leave a comment

Allowed HTML: <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre>

Photo of Bruce Schneier by Per Ervland.

Schneier on Security is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc..