Schneier on Security
A blog covering security and security technology.
« "Protecting Europe Against Large-Scale Cyber-Attacks" |
| Privacy and Control »
April 5, 2010
Detecting Being Watched
This seems like science fiction to me:
The camera uses the same "red eye" effect of from camera flashes to project it hundreds of meters, allowing it to identify binoculars, sniper scopes, cameras and even human eyeballs that are staring at you....
Posted on April 5, 2010 at 1:30 PM
• 51 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
Even if it does work it's going to be generating all kinds of false positives from people glancing at you and random glass reflections.
Couple that with sensors tuned to the wavelength in use and you'll be attracting watchers all around :D
Plausible provided that they're using specific wavelength absorption to distinguish eye-reflection from random backscatter and reflection, but the coatings and optics in binoculars and gunsights make anything but naked-eye detection doubtful.
Red-eye effect detection is hot in research in that it helps locate features on a facial image and allow scaling but wearing glasses and how a person faces can affect it badly.
Given that Xerox holds one of the original patents
and that the idea is used in many cameras' "Face finder" algorithms, it's not surprising that they'd come up with this.
The real question is if this is original or if it's nothing more than a good digital point and shoot camera plus IR emitter in a fancy box hiding behind a "Performance and instrument specifications are 'need to know' and require a disclosure agreement in place" statement, and a Pentagon-sized price tag.
I recall a long time reading about a device simiar to night-vision goggles which exposed hidden cameras by putting out IR light and then showing where it was being reflected from; supposedly the lenses of cameras reflect the IR back, and would show up as shining dots on the device.
Maybe this is something similar?
IIRC Back before the wall fell the CIA warned the DoD that the next generation of Soviet tank (FST) would have a system something like this to detect reflection from ATGM weapon sights along the line of advance. The neatest feature of this purported system was that it would then use a coaxial "not eye safe" laser to dazzle/blind anyone taking a bead on the protected vehicle...phzzzt! One supposes there are significant IFF and false positive issues that arise from such a safety solution. Anyone else miss the Cold War?
In certain places and at certain times, it might be unlikely enough to have anyone looking at you that this kind of gadget would actually be useful, assuming that the time taken to know that someone is looking at you doesn't distract you from getting killed by someone else...
But if there's a narrow range of wavelengths associated with this pseudo-redeye effect, the countermeasure seems pretty damn obvious. And the detector itself tells you where your target is.
We tried this back in 1992 to detect whether someone at an ATM was wearing a motorcycle helmet, and even filed a UK patent application. Perhaps we gave up too soon, but my instinct is to agree with the sceptics: it might be hard to discriminate well enough for a decent error rate.
"Do not look into bezels"
"'Do not look into bezels'
Probably an IR source that's strong enough to be a problem if you're close enough to read the label.
I'm guessing it's clouded by some marketing speak. There have been visible laser scan devices to find hidden cameras by using reflections from lenses for a few years now. For example:
My guess is that this is based on the same principle but uses a higher power non-visible light (UV? IR?) laser.
@mcb: Puts a whole new meaning on 'Explosive Runes'.
Uh, Bruce, this isn't really new. Perhaps an evolution of an old concept, but people have been doing this with scopes and pinhole cameras for a long time. Here's a top flight bug sweep company's equipment for using infrared emissions plus camera to spot covert cameras (among other things).
TESA Infrared Eavesdropping Detection Technology
The pictures on their site that show each piece of equipment at work are pretty interesting. The description of Jet Protect's tool just sounds like a more portable, less assured, version of TESA.
This is something the DoD has been looking into for many years -- looks like someone finally made a usable product. In the asymmetric warfare we tend to get into these days, finding a sniper before he shoots you is REALLY important to be able to do.
At one point, our guys also looked at making one powerful enough to blind the guy -- but decided against it.
Sure there would be false positives -- but helping you just re-scan the 20 false ones, instead of a detailed long term look at 1000 possible hides is a big improvement. When that bush or pile of rubble has a big eye in it looking at you -- well, doesn't take much more checking out.
Good thing you don't need this fancy expensive stuff to detect being watched in the domestic surveillance programs here on US soil. Just look in your rear view mirror and go around the block, twice.
Been done for years as anti-sniper technology - you just look for retro-reflections from the scope. There are even anti-anti-sniper countermeasures, a beehive hood = a set of hexagonal tubes on the front of the scope to limit the angle you can see the lens from.
False positives are less of a problem, there aren't many optical elements pointing at you from a forest on a battle field.
Not going to work a damn on the street though!
Countermeasure: Metalized balloons. Should play havoc with it.
I love the "do not look" warning. Reminds me of the title of George Lakoff's book "Don't Think of an Elephant". Can there be a Streisand Effect for product safety warnings?
Snipers I knew often favored plain "iron" sights
Annoyingly the source material is blocked by my corporate firewall as "violent / offensive material" so I cant get to see it. This means any comments I make are based on previous examples of this sort of technology, so forgive me if I am totally off track in places.
However, in general, I am unconvinced by this sort of technology and should it work, it seems fairly easy to defeat.
Reflections from a camera are an odd one - what is expected to cause the reflection? Different camera systems all have different focus / image capture methods (SLR vs Video vs Mirrorless etc). Is the return signature caused by reflection off the glass alone? (Madness).
Same with scopes and the like. Most of the optics are geared to reflect light towards the eye, not out of the front, so getting that return signature wont be easy.
The red eye effect from cameras is inconsistent at best (and related to proximity between flash and lens) and can be removed by simply changing the angle the flash light hits the eye. Previous systems of this nature seem to be easily defeated by not using the central eye focus point to look at the object and this is something we do on a daily basis.
If it doesn't work - you either don't know by virtue of not detecting the watcher, or as they put it - you get shot - and then you can't complain..
My first thought for evaluating it is - is it restricted under ITAR. Can't find anything 'bout that on the site.
See what happens Bruce post something about binocular detection, imaging - get pr0n spam to help us with our loneliness.
What if You put absorbing IR filter (available on SLR cameras) in front of your lens?
Or just look into reflecting IR filters, and choose one that will reflect in angle other than arriving.
I suspect the obvious method of filtering most false positives is checking source stability, i.e. scopes move (even a little), office building windows do not.
soldiers will start carring bags of marbles glued in pairs by strings to make eyeball signatures to lure snipers into revealing thier positions by shooting
I drew explosive runes today.
"This seems like science fiction to me"
What was it Arther C Clark said about science and magic?
You realy surprise me though, it's not as if you have not blogged about systems that work a similar way (the "anti-camera system") and had the technology explained in quite some depth, by myself and others.
Nor is it the first time I have mentioned "TEMPEST's dirty little secret" that all monitoring systems can not only be found by the equivalent of "red eye" but enumerated as well...
Agh well just to go through it again, this is not particularly new technology in fact parts of it go back into the early 1980's if not earlier to my certain knowledge (I was involved with the design of a prototype 3D camera for inspection work in nuke reactors).
And if these people have a patent on the system then they or the patent clerk have failed to be "duly diligent" in searching for prior knowledge.
The trick behind this sort of system is to work out how to do it reliably and it's fairly simple when you remember how John Logi-Bairds "flying spot" TV system worked.
Pick one or more EM frequencies that have a low background level in your chosen environment but have a high return for the optics you are looking for.
Try and make it a coherant source if possible and use a colimnating system with in built helical scanning (two orthagonal prisms rotating on their axis at appropriate intiger ratio).
You now have your "flying spot", you now need a receiver for your chosen EM frequencies. By using a half silvered mirror in the right part of your scanning system you can pick up the reflection with little difficulty (and avoid quite a few other problems as well).
If you also pulse your coherant EM source very rapidly with something like an orthaganal Gold / JPL Ranging Code you can avoid interferance problems as well as being able to reasonably reliably give range (tactical Radar systems use a similar technique).
You can tell what is looking at you in a number of ways.
First of the human eye twitches rapidly as part of the way it works, this is to do with the way our nerves transmit information to the brain. If the eye did not twitch your eye sight would "white out" (you can try an experiment to show this with a ping pong ball with a hole in it over your eye and coloured lights).
This twitching as you would expect modulates the return signal with all sorts of interesting information (research is currently underway to use this for non-invasive medical diagnostic treatment amongst other things).
Oh and unless it's an "infinate focus" optic you can also tell if it is actualy "focused" on you or not (which applies to the human eye and some high gain optics as well like spotter scopes)
The other thing is to use a "tunable coherant" source and sweep across various EM frequencies, by the simple process of absorbtion you can very much tell what is at the focal point of the optics you are trying to enumerate, and in many cases the optics that preceade it.
So is it reflecting just off/internally to the lens, or through the full optics path (ie back of the sensor)?
My guess is that they are specifically looking into things that are retro-reflective. Reflections off of shiny objects or windows look nothing like red-eyes or the IR reflection you get off of a digital camera. (Digital cameras have a "Hot Mirror" specifically to reflect IR light back, away from the sensor.)
So this product sounds distinctly plausible to me, though the devil is always in the details.
As far as counter-measures, I would use Illuminite.
Perhaps only the preceding commenter (SalusaSecondus) understands the concept of this gadget. For example, one comment above suggests that the anti-reflective coatings on (for example) lenses of rifle sights would reduce the effectiveness of such a system.
Not so! Anti-reflective coatings help it to work better. If this gadget works the way I think it does -- and if it doesn't, I doubt that it could work at all -- it does NOT look for specular reflection (sometimes called glint), such as might be seen on lenses, window panes, beer bottles, etc.
Those of us who drive at night occasionally see the eyes of a critter, on or near the road, looking at our approaching vehicle. The eyes appear MUCH brighter than any nearby object -- BUT ONLY IF they are looking in the direction of the vehicle.
The distinct bright patch that can be seen when shining a light directly into an eye or camera (that is aimed toward the light source) is not a reflection off the cornea, camera lens, etc., but a rather reflection from the imaging element itself (retina, CCD, etc.). Either the imaging element is in the focal plane, or it can't make a clear image. And if it is in the focal plane, light reflected from it must travel back through the focusing optics in the direction from which it came (reciprocity). It works like the optical procedure of autocollimation.
It is common to manufacture structures as retroreflectors (reflecting light in the direction from which the light came) -- for example, street signs, reflectors on the rear of vehicles, reflective safety patches on clothing, etc. These are intended to work over a broad range of angles. Eyes and cameras are retroreflectors that work in a small angular range: only near the optical axis. I suppose that manufactured retroreflectors would be the biggest challenge to a detector like the one described here.
SalusaSecondus says: "Digital cameras have a "Hot Mirror" specifically to reflect IR light back, away from the sensor."
Well, they *might* have that. But many cheaper cameras do not. I long ago learned that I could detect 'bright' infrared pulses (such as from remote controls, or from the newer smart electric meters) by simply checking through my cell phone's digital camera.
It would a little embarrassing if your super-high-priced "viewing detector" could find super-high-priced imaging devices, but couldn't find a cell phone camera left in a non-obvious location with it's video recorder on.
It does appear that the fundamentals behind this are similar to this device ... http://www.brickhousesecurity.com/sf-103.html
@Chris S "could detect 'bright' infrared pulses "
Yep it do...cool. Thanks Chris.
I remember reading about a similar device years ago for detecting bootleggers in movie theaters. Apparently there's a market for crude video recordings of newly released movies....
"So is it reflecting just off/internally to the lens, or through the full optics path (ie back of the sensor)?"
I'm not sure at which end you are talking about. So,
The full path,
Starting at the EM source:
1,Coherent EM Source,
2,Beam spread collimator,
3,Half silvered mirror (directional coupler),
4,Beam reduce collimator,
7,Primary lens structure (to give wider field of view/coverage).
Outbound into free space,
Then at the observer/snipers end (ie a refracting telescope):
A,Primary/Objective lens structure,
B,Focal lens beam spread collimator,
D,Human eye lens,
F,Photoreceptors (rods/cones) in retina.
Now if all of this is in focus (ie observer/sniper is actually looking at the target the EM source is mounted on) you get the red eye effect. That is the coherant EM source will be focused onto the retina and will reflect back through 180 degres (give or take a fraction) and return along the same path but in reverse untill it meets the half silvered mirror that is used as the directional coupler where the return path beam is directed through another collimator and onto your photo detector array.
This photo detector array (CCD detector) is where all the good stuff is done with optical filters etc to obtain the maximum optical information, before the relavant signals hit the DSP's to extract other information.
For instance with more than a single frequency EM source, it is possible to enumerate (detect) the different substances at the focal point where the 180 degree reflection takes place.
Thus you can tell not only if it's a silicon / galium / human retina as the sensor, it is possible to "image" which part of the retina is being used (rods and cones have different reflection spectra).
Also it should be possible to measure the response time etc of the eye which is supposed to be not just a secondary (to the retina) "unique bio-metric", but potentialy an "at a distance medical diagnostic" indicator of heart rate, blood oxygenation, core body temprature, stress levels etc etc.
Although I'm somewhat doubtfull if there would be sufficient signal to noise on some of the medical diagnostic signals at any significant range (say no more than a couple of hundred feet tops).
The reason I mention the medical diagnostics is there have been some claims (for research funding) that the medical diagnostic potential of some "at a distance retina scanners" can be used as duress / lie detectors, as well as a method of determaning "ill health" in airports etc to detect the likes of SARS etc...
And thus it is (signal to noise permitting) a logical follow on in the technology also I mention it to scupper any non US potential patent claims as it is now known and thus "prior art" (sadly in the US there are "submarine patents" which I have yet to see the good of so I expect this to be hitting a US patent clerk any time real soon...).
@ BF Skinner,
"Yep it do...cool. Thanks Chris."
As I've mentioned before, those phone cameras are also not just very good at detecting IR security lighting but covert CCTV cameras as well...
If you get one of those LED torches/flashlights and take out the "clearwater LED" and replace it with an IR LED (from a TV remote control). You now have an IR EM source...
If you hold the IR source very close to the phones camera (but pointing away like the flash) you will see the "red eye" effect of most CCTV cameras on your phone screen, and you can also take a picture of it as evidence to the fact...
All "monitoring" equipment has failings (belive it or not but you can "hear microphones" if you know how to do it).
They usually have one or more of the following failings,
2, resonance (/ antiresonance),
And the how and the where alow the device to be enumerated (and even Quantum Crypto has fallen foul of this problem, making it decidedly insecure).
So this requires that the "refracting telescope" is pointing at the detector. Seems that operationally this is not quite as useful as it sounds.
However counter measures if full spectrum is used would be hard, and expensive. Not impossible.
above was by me @clive... i had a brain fart...
Going to the coffee machine now ;)
for my own education:
The observer has a SLR (digital or otherwise) mounted on a tripod and has used the viewfinder to manually focus on the target with a remote shutter release to take pictures. The mirror is down covering the film / sensor.
What part of the camera system is detected by this device?
Is it only capable of detecting the eye of the observer?
Can it still detect the observer when he is looking through the viewfinder?
If someone is using a film video camera, what part of the system is detected by this device?
Can you expand on the Gold code modulation idea. It would seem like a lot of bother just to range find, but obviously you should be able use it improve the signal to noise ratio of the system, especially for very long spreading codes. Presumably you can also track any optical path length modulation (I'm thinking of heat bending the light beam)
My minds working overtime trying to figure out what application needs this functionality...
Gold codes etc are very light in hardware and easy to use for range finding. I used them with sonar with a $5 microcontroler and was getting >5mm resolution at 500m without even trying.
Optically its also easy to use and would give better results that a simple pulse and doesn't need silly accuracy requirements on the clock jitter.
If you are modulating the signal somehow, what alternatives would you suggest? Gold codes and m-sequences are the out of the box standard for this sort of thing.
More abstractly they are effectively just a "modulated" pulse with a matched filter, and a chirped pulse can be used with simpler hardware. This is/was done on radar systems. A lot of the hardware can then be done with SAW devices. optically one could also use chirped pulsed methods as well.
I'm very familiar with gold codes, but I could not understand why the application benefit from highly accurate range finding, I've thought about it a little more and can see where this functionality would be very useful in an open battle field, possible combined with an automated system to collect the exact coordinates of a suspected sniper positions...
In addition to gregs points the orthaganality of such systems have advantages when several systems are used together and their "view fields" overlap. As would be quite likely in a convoy etc where vehicals had independant systems.
You also need to think about other aspects such as "offset LIDAR" etc.
Although Gold codes are very good they can be slow to "lockup" when starting from cold on an unknown long sequence.
So also have a think about NASA JPL (Jet Propulsion Labs) Ranging codes, they have similar orthaganality but as they are made of short Gold codes can lock up in much much shorter time periods.
The question of,
"I'm thinking of heat bending the light beam"
Is a curious one and I'm assuming you have some knowledge of heat lenses.
More pragmaticaly however it is not unknown for ranging systems to use two coherant but orthagonaly polarised and modulated EM sources. The reasons are a bit complicated but if you have a read up on weather lidar/radar it will give you a starting point as to the physics (have a trawl therough the optinfobase on lidar for some of the details).
Put simply you can measure wind direction and other things of interest for "auto gun laying", which if you think about it is the next stage of dealing with an identified hostile.
Agh you posted whilst I was responding, and I see you get the auto gun laying potential.
I hope the coffee has done the job (I've yet to have mine and it's already a quater past afternoon tea time here in London 8)
"So this requires that the "refracting telescope" is pointing at the detector. Seems that operationally this is not quite as useful as it sounds."
It does not need somebody to be using a sniper scope, a camera or other optics including the human MK1 eyeball will do.
Importantly it tells you you are a "point of interest" to somebody.
Think of say finding a UAV by lidar and determining if it's having a shufty at you.
Or an example is "smart weapons" where forward ground troops "paint the target" only at the last few seconds before the bomb hits the designated target.
All the systems I've seen use a pasive telescope to "spot" the target prior and during the "paint".
With a fast auto gun lay you could take out the spot system before you as the target get painted, and thus the smart weapon would just be a traditional gravity bomb and probably miss by a long way.
Oh and of course you could also use the system to locate the smart weapon in the air and pump a very high energy light pulse into it's sensitive optics to blind it and it again becomes a traditional gravity bomb.
And as for those Apache Gun Ships, they might not be to happy to be spoted and have high energy light pulses followed by missiles sent in their direction as soon as they start looking around, it would kind of take a significant amount of their advantage away.
In modern conflict the soldier and the pilot are not realy the ones making the targeting choices anylonger if they want to survive, they are just not quick enough.
Sadly if you think about it this sort of system makes blue on blue more not less likely 8(
A typical SLR camera would presumably NOT be detected, unless either (a) someone is looking through the viewfinder, or (b) the shutter is open. The reflective surface needed to "complete the circuit" is either the photographer's retina, or the film/sensor.
In this sense, the SLR must actually be "looking" (actively receiving image information), not merely pointed, in order to be detected.
Interestingly, if the SLR has a focal-plane shutter set for quick exposure, only a little of the imaging surface would be exposed at any time during the interval of exposure. If the shutter is much less reflective (I believe this would usually be the case), a focal-plane shutter SLR could return a comparatively weak signal during exposure.
Thanks I was not thinking of multiple overlapping FOV's but clearly this would be desirable.
On the gold code lock problem, I was assuming that code synchronization would be done optically using a Kerr or Pockel cell, but I guess you don't need to modulate the source that quickly, so a simple digital system lets you easily correct for Doppler effects and should also enable you to tell the refractive index of all the elements in the optical path, which could be handy for an automated fire system.
I'm still not so sure that there is an advantage to ranging this way vs a simple OTDR, I guess the best solution depends on which technology you are more comfortable with.
"I was assuming that code synchronization would be done optically using a Kerr or Pockel cell, but I guess you don't need to modulate the source that quickly"
That's a point to you on the technology side, now I understand why you said,
"My minds working overtime trying to figure out what application needs this functionality..."
We're thinking more than a few orders of magnitude different in range resolution capability.
I'll put my hand up and say I had not even considered this way, simply because the last time I looked the technology was a tads expensive and fragile.
But I would definatly be interested in hearing your guestimate on (uncased) BOM on your system, as technology "sweet spots" are always on the move.
This is because despite what various Politicos say to the press and House of Commons Select commities, the bitter truth is the likes of the UK MOD procurment people are looking for "as cheap as chips" throw away technology these days.
All delivered to the battle front by the likes of DHL/TNT/UPS (belive it or not, tanks and other heavy armor where being transported by drivers working for "Blue Arrow" not so long ago).
Thus I'd be the first to admit the system I was thinking of would use a very cheap but robust optical head, that with say a very minor change in the case and primary lens could also have a myriad of uses such as civil engineering. The back end being all software on COTS hardware. Think about the history of GPS and what sort of GPS equipment squadies get issued with these days to see why I'm think that way.
As you say,
"I guess the best solution depends on which technology you are more comfortable with."
A little history explains my sometimes odd outlook ;)
I moved out of MilSpec design quite a while ago as I could see the writing on the wall. The wake up call was at a retirment party, the retiring person told me that they had spent a quater of a century as a projects manager on just one MilSpec project. And I can clearly remember thinking about how the likes of Intel and AMD where slugging it out on 18month product lifetimes.
It was clearly not sustainable and it's what made me get out of that sort of design. Via an odd route I ended up doing Fast Moving Consumer Electronics (FMCE) design for personal communication (phones amongst other things).
But even there I saw the writing on the wall with even shorter product design and life cycles.
And again it came back to me with the first Gulf war and failing comms equipment and tentative enquiries for MilSpec or Hardend mobile phone systems. Needless to say the "old guard" won out. But look at the reality of today, squadies find they get better coverage on the battle field with their personal mobile than they do with their "MilSpec Comms equipment". Likewise the UK's Met Police with their comms systems.
The joke of it is the longest hardware project life cycles these days are to be found in the Civil airframe, Telco Infrastructure and civil space technology industries. Where the recent economic slump has started people taking a good long hard look at cutting upgrade cycles to about a half what they where, thus nearly trippling life expectancy and availability requirments, there's 747's flying comercialy with something like 30+ years on the airframe, and no plans to retire them.
It's why I'm looking at the likes of Software Defined Radio (SDR) on consumer grade electronics in high availability configs. I suspect it's the way it's got to go to keep up with expectations.
Oh and of course as "software for everything" applies to critical infrastructure like mobile phone base stations, we need real software security right down through the OS etc and below layer 0 on the software stack and well into digital hardware that would have been the domain of analoge electronics just ten years ago.
It's why Nick P is looking at "secure by design" CPU's for the likes of I/O control, and I'm looking a little further into the future by thinking on how to put multiple "prisons cells" on the chip. Just to protect the RF amp, Down Convert, DSP control systems on that Government Mandated box that controls your home energy usage for the next thirty years...
@ MarkH at April 7, 2010 3:15 PM
Thanks for the information.
Some additional questions then (although I am already in the process of writing this off if it cant detect the camera except for the 1/200th of a second its taking a picture..) if you dont mind:
When the observer is looking through the viewfinder, is there enough of a return lightpath for it to generate a detectable signal?
Would a live-view system on a DSLR combat this? (in my ignorance I have no idea if live view lifts the shutter or not).
If it was a film SLR, wouldnt this mean there was no return signature to pick up even when the shutter was open?
- - -
In general terms then, am I right in thinking that this is a system that can detect a camera only in the act of taking a picture?
So if its bright daylight and the film/sensor has a good ISO, you may have to be checking in the right direction, at the right angle during the 1/8000th of a second the shutter is open?
"In general terms then, am I right in thinking that this is a system that can detect a camera only in the act of taking a picture?"
Depends what you mean by "taking a picture".
For an old fashioned film camera you have a shutter that would in most cases stop the 180 degree reflection as the focal point is interupted (however there is a reason why the lens side of the shutter is painted mat black and that's multiple reflections).
However the view finder is still acting as a "telescope" so when the eye is behind the eye piece the retina will be the 180 degree reflecting point.
So film cameras are the same as real SLR's.
However, most digital camera's including most Digital SLR's have a sensor always in the optical path at (one of) the focal point(s), so nearly all digital cameras will be detected when pointing at the detector.
As I've indicated this is not exactly a new idea, and Bruce actual bloged about a system to stop photos being taken at exhibitions and trade fairs which not only detects the camera but in the case of film cameras can actually fire a bright light into the camera optics effectivly blinding it with lens flare etc.
As I said at the time it would be interesting to see what such a system would make of "clear water" LED's (as found in most LED flashlights) which are effectivly a clear lens with a lump of polished silicon at the focal point...
Thanks for getting back on this.
My understanding is:
In an SLR the light path through the lens to the pentaprism is almost the same as from the lens to the film / sensor with a mirror in the way to allow you to look through the lens. When you release the shutter the mirror flips up and the aperture is used to control the light hitting the film / sensor.
Now, with the mirror down and no eye looking into the viewfinder, what would this system detect? The light path would just go through the lens, hit the mirror and out of the viewfinder.
I am sure you can recall time spent observing - very little of it is done actually looking through the camera - you line it up where you want it and then basically hang around keeping an "eye" for something to happen.
As I see it, this system would only detect the observers eyeball while s/he is looking at the detector or looking through the camera at the detector. While it is possible that you would scan the right place at the right time, its a low probability. Even then it only works if it can generate enough of a return signature along this convoluted light path.
What can you cover your windows with that will Prevent an Infrared-pointing device from listening to you illegally?
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.