New System to Counter Photo and Video Devices 401
Incongruity writes "News.com is reporting that a team from Georgia Tech has developed and demoed a system that actively searches for and effectively blinds cameras and camcorders within a 10 meter radius." From the article: "In this system, a device bathes the region in front of it with infrared light. When an intense retroreflection indicates the presence of a digital camera lens, the device then fires a localized beam of light directly at that point. Thus, the picture gets washed out."
Re:I thought the same thing... (Score:5, Informative)
Re:I thought the same thing... (Score:2, Informative)
It isn't the IR that blinds the camera - the IR is simply to look for reflections, like those given off by lenses (and, of course, eyeglasses). When the IR gets a "hit", a directed beam of light (flashlight on a servo?) is aimed at the lens. Pretty low tech really. Given that we've had "lens detection" devices for years (decades? The military is a big fan), the real story seems to be the rather lame application. I guess the "amorous couple" (per the article) is going to carry around some sort of detection/light device with them? Give me a break.
Lame.
Re:What about (Score:5, Informative)
How it works
The Georgia Tech system essentially exploits the "retroreflective" property of digital camera lenses. When light strikes a retroreflective surface, a portion of the light bounces back to the original source. While eyeglasses, bottles, watches and other glass surfaces are retroreflective, a coating on virtually all digital camera lenses puts cameras in a class of their own.
People too. (Score:5, Informative)
People, too.
That's why you get "red eye" in the picture if the flash is too close to the lens.
For people it's probably a vestigial remmanant.
For animals it's a night-vision adaptation. The retro-reflector is behind the light-sensitive part of the retina. Any light that makes it through the sensors is sent back (nearly) the way it came in, giving the retina a second chance to catch it and thus a tad under a 3db increase in sensitivity - at a slight cost to focus. The shine you see is what made two passes without being caught.
Re:Infrared filters! (Score:3, Informative)
TFA mentions that IR filters "present some challenges ... though it turns out that the camera detector can spot lenses cloaked with infrared filters." However, I think if someone intentionally tried to secure a camera against this device, they would have a lot more luck. The filters they tested against probably had similar glare properties to a camera lens. So take a camera with an IR filter, give the filter itself a good non-glare coating, put something like the Leopold Anti-Reflective Device [outdoorsuperstore.com] on it, and a nice lens hood/shade to keep it from catching any periphrial light. If the device still detect the camera without its sensitivity being turned up so much that it blinds everybody with glasses, then I will be a lot more impressed with their technology.
You're wrong. (Score:2, Informative)
Re:FP? (Score:1, Informative)
To circumvent this, I would try: (1) a convext IR-cut filter, plus postprocessing to minimize distortion, and (2) a shirt covered with flat IR-reflective objects, on the assumption that the system can only shoot a certain number of light beams per second.
Re:What about (Score:5, Informative)
A group of cameras are arranged in a ring formation, with their lenses facing inward. Typically, this ring is raised up about 10' or so above the ground, and the cameras aimed down toward a common area. Each camera's lens has a donut shaped ring mounted to it. The donut's surface is covered either red or ir emitting led. The light from these leds floods the capture area (the volume) and bounces off of reflective markers which are attached to the actors inside the volume. The cameras, which are IR sensitive pick up the markers, and a computer then uses the feeds from multiple cameras to triangulate the positions of the markers.
Anyhow, the Vicon guy did say that its not a good idea to stare into the strobes, as it was probably not healthy for the eyes. The red ones are probably less unhealthy, as your pupils contract due to the visible light. The ir ones don't emit any visible light, and the only way to tell if a strobe as working was by a green led stuck among the ir ones.
Just to wrap up this mishmash of info, and to make a point, I don't think flooding areas with ir light is a good idea, as I did find myself getting headaches and eyestrain if we left the strobes running too long in the studio.
Re:People too. (Score:1, Informative)
Some facts from their Research Paper (Score:2, Informative)
Seems it can be "tuned" to detect the retro-reflectivity of CCD Cameras. Also, the authors did say that there are many ways the system can be fooled. Personally, I would just attach a paper tube to the camera, long enough to allow a photo to be taken while blocking out the IR beam from the detector. For those worried about getting IR beams in their eyes, remember that they are just using your standard IR LEDs, not LASER LEDs. From the paper: In summary: 1) It is harmless if a false positive (camera-like device) is detected. It cannot damage cameras. 2) This probably won't work on CMOS cameras, which are likely to be the next generation technology used in digital cameras. 3) Limited angle of detection, range of detection (based on resolution of sensing camera) and numerous counter measures makes this system an interesting prototype at best. Its still a ways off being used by the 3-letter government agencies.
Re:I thought the same thing... (Score:2, Informative)
RTFA, the infrared is only used to detect the camera lens. It's not used to blind the camera. A normal project is used to blind the camera. So you won't go blind without knowing what's hitting you. A valid question would have been, "wont it think my glasses are camera lens". The answer to which is also provided in the article. Again... RTFA.
Re:Lawsuits here we come! (Score:2, Informative)
This method, however, has severe limitations:
1) It relies on the CCD sensor being visible through the lens. That means that SLR cameras, for example, cannot be detected, since the mirror is in the way. For that matter, any camera that uses a closed shutter that only opens when the photo is being taken can't be detected. This does rule out any camera with live preview, however.
2) A simple IR filter in front of the sensor will also defeat this system. As long as the filter will absorb IR, instead of reflecting it, the system will not be able to detect it. Take a cheap Vivitar or Concord cam, and glue a $30 IR filter on the front lens element, and you have a $100 compact cam that can't be detected.
3) Point your lens somewhere else. Start capturing a movie. Point the lens at the target. You will still get a few usable frames before the system detects your camera and blinds it. One frame is all you need.
In short, there's plenty of way of defeating this, and anyone who's even remotely determined to circumvent it will.
Re:I can just see it now... (Score:2, Informative)
Don't believe me? Check the extremecctv.com web site.
Also, thinking of blinding the camera with IR? I had a product idea so I tried it, thinking that abberations/imperfections in the lens would be enough to confuse the per-pixel DSP capabilities of the camera. Bzzt. Wrong. Even an LED array surrounding your license plate won't blind Extreme's cameras. Pelcos, Toshibas, and most Sanyos? Yes, but Extreme designs special ops cameras for the military, cameras for missile guidance systems, and other uber-high end systems, and they've transferred some of that capability to revenue generating, er, I mean law enforcement technology.
I know I sugested that about a year ago on /. (Score:5, Informative)
Part of "No Pictures, Thanks" from 1/26/05
It's actualy easier, you just need a high powered IR source, such as a bunch of LED's,
the Cameras AGC automaticly adjusts so you turn totaly dark.
Re:I can just see it now... (Score:1, Informative)
digital SLRs, like the Canon EOS 1Ds Mark II [canon.com] or Nikon D2x [nikonusa.com] in order to get their shots to a magazine/paper/outlet, quickly. The Canon's 16.7 / Nikon's 12.7 megapixels is good enough for even fine art magazine shooting.
Re:You're wrong. (Score:3, Informative)
So why not use the infrared filter to prevent the detection of the CCD camera. Don't reflect the infrared light back to the detection device. Thus no camera detected and no visible light sent to wash out said camera.
Re:I can just see it now... [OT] (Score:5, Informative)
In my (not so) humble opinion the law should treat tail-gaters as harshly as drunk drivers. There's no excuse for either and both are incredibly dangerous to other road users.
</rant></offtopic>
Re:I thought the same thing... (Score:4, Informative)
Actually, they shouldn't. It is against the Geneva convention to use devices such as lasers to cause blindness. Death's okay. But blindness is verboten.