Resolving power

[Sheki]

For quite some time now, whenever confronted with stories about the "evil gumint and their spy-satelites that watch you while you pee", I would scoff and ridicule the proponents of such ideas. You see, having perused the BABB for several years I had become well aware that the resolution required to recognize facial features, read text, etc, etc from space, would require a "ginormous" orbiting telescope - and hence was just not feasible. But then last night it occurred to me that this is not necessarily true.

I am thinking there are two key technologies that if used concurrently could be used to achieve pretty fantastic resolution:

1. Adaptive optics
http://www.universetoday.com/am/publ...on.html?332004

Basically, I see no reason why a space - based telescope could not be fitted with adaptive optics (employing the "laser as guidestar" method) to peer down through the atmosphere, as opposed to a ground based telescoping peering up through said atmosphere.

2. Space-based Interferometry
http://arjournals.annualreviews.org/...fb?cookieSet=1

and

http://www.coseti.org/topic16.htm


In short, I am assuming that it might be possible to construct "keyhole"-type satellites that employ both adaptive optics, and are capable of networking to perform optical interferometry.

I am by no means suggesting that such a system currently exists, merely that I can no longer scoff at the idea that a satellite might be reading my licence plate.

So my question is, have I lost it? Is all of this remotely possible? Any ideas what the theoretical resolving power of such a system might be? Comments?

Sheki

[Kristophe]

Adaptive optics simply allow a telescope to achieve an operating resolution closer to its theoretical maximum. The resolution won't be greater than lambda/d, where lambda is the wavelength of light, and d is the diameter of the telescope.

An interferometer isn't going to be able to read your license plate, especally if your car is moving. Not unless you have a number of telescopes set up in the right configuration. The resolution of an interferometer is lambda/D, where D is the distance between the two telescopes, but only along the line that connects them. Assuming they're arranged in a horizontal line, then the resolution is improved horizontally, but not vertically.

[Sheki]

Adaptive optics simply allow a telescope to achieve an operating resolution closer to its theoretical maximum. The resolution won't be greater than lambda/d, where lambda is the wavelength of light, and d is the diameter of the telescope.

An interferometer isn't going to be able to read your license plate, especally if your car is moving. Not unless you have a number of telescopes set up in the right configuration. The resolution of an interferometer is lambda/D, where D is the distance between the two telescopes, but only along the line that connects them. Assuming they're arranged in a horizontal line, then the resolution is improved horizontally, but not vertically.

Yes exactly. With enough keyholes you could have quite a few baselines of various lengths and orientations. I am picturing a form of interferometry that combines multiple baselines. Configure the satelites in a circle (or other pattern) and then play connect the dots with the baselines.

Sheki

[A Thousand Pardons]

Is all of this remotely possible? Any ideas what the theoretical resolving power of such a system might be?

As Kristophe said, your number 1 doesn't contribute that much--sure there is an improvement, but that's still irrespective of the theoretical limit, because that's based upon aperture. With number 2, there is no theoretical limit on aperture, but the cost is enormous. Probably.

[joema]

KH-11/12 optical spy satellites are somewhat larger and heavier than the Hubble Space Telescope. They are about 4.5 meters in diameter, 15 meters long, and weigh about 18 tons (36,000 pounds, or 16,300 kg). About 14,000 pounds of this is maneuvering propellant

The optics are thought to be about 2.5 meters in diameter, roughly the size of Hubble's primary mirror. They likely have adaptive optics of some type.

The orbit perigee is variously reported as 95 miles (153 km) or 170 miles (274 km). Given the fuel capacity, it's likely they can maneuver substantially.

http://www.fas.org/spp/military/program/imint/kh-12.htm
http://www.astronautix.com/craft/kh12.htm

From this we can calculate the theoretical linear resolution. Angular resolution at 500 nanometers is:

resolution = 250000 * wavelength / diameter, or 0.05 arc seconds

Linear resolution at distances of 153 and 274 km is given by:

res = tan (angle) * distance, or 3.7 cm (1.4 in) at 153 km and 6.6 cm (2.6 in) at 274 km.

Even assuming perfect adaptive optics they can't read license plates, but they can see rough human body shapes.

The Lacrosse/Onyx radar-based recon sats use synthetic aperture radar to image though clouds. Their resolution might be about 1 meter:

http://www.fas.org/spp/military/prog.../xlacrosse.htm

Although some speculate about synthetic aperture optical interferometry using single optics, this seems extremely unlikely. Even use of conventional optical interferometry (outrigger telescopes) seems unlikely. It is extremely difficult to make that work on the ground, much less in space.

A further challenge is reconsats (unlike Hubble) have a fast-moving target. At a 160 km perigee, a terrestrial target moves at about 3 degrees per second. Any optics must somehow pivot to compensate for that motion. Some speculate the KH11/12s have tracking heliostats to avoid constant vehicle motion.

After about 15 more years of development, NASA's Terrestrial Planet Finder will use space-based infrared (not visual spectrum) optical nulling interferometry to image extrasolar planets: http://planetquest.jpl.nasa.gov/TPF/tpf_index.html

[Sheki]

Thanks for that joema. That is exactly what I was looking for.

However, one could argue that what NASA is developing presently with the terrestrial planet finder, the military may already have accomplished. Afterall, was this not the case with adaptive optics? (ie. the military had developed adaptive optics for "other" purposes, and the technology was later adapted to astronomy). Or is that just an old conspiracy theory...?

Sheki

[joema]

It's always possible they are a few years ahead in certain limited areas. However their main need is not more optical resolving power, but better coverage, esp during cloudy weather. What's the point of having 1 cm resolution if your primary targets are often cloudy?

Also the satellites are in polar orbits that give about two useful passes per day over a given target, assuming good weather. They pass from horizon to horizon in just a few minutes, and their orbits are well known. After that there's no coverage.

The real need is more satellites and better all weather coverage, not higher resolution.

Synthetic aperture optical interferometery using a single orbital telescope is science fiction IMO. It works for radio frequencies, but not visual optical wavelengths.

They'd do better to build a huge geostationary optical recon sat at 22,500 miles, and use outrigger telescopes (possibly free flying) to improve resolution. But to provide equivalent resolution at that altitude would require a 250 meter diameter telescope (or equivalent). That's immensely challenging but at least it's barely conceivable.

[Kaptain K]

The "problem" (or one of them) with opical interferometry is that the separation between elements must be known (and constant) to within a fraction of the wavelength being used. This is relatively easy (but not trivial) with radio, where the wavelengths are measured centimeters or meters. For light, you need known and constant separations measured in 10s of nanometers.

[the_shaggy_one]

The government may not be able to use a sattelite to read your license plate (or watch you pee), but they sure could use one of the many spyplanes they have. I think the U-2 is still flying around up there, as well as the new UAVs. Arm either of those with IR cpability (both have it already, I'm sure) and the government could probably see the residual heat on your toilet as you're peeing.

[Russ]

It is not my intention to make y'all parinoid but...during my all expense paid vacation to the jewel of south east asia, G-2 would provide us with photos of our objectives that were stamped "Sat. Com. Cen., G-2 sec.", etc.

That's military speak for Satelite Command Center, Group 2 (intellegence) section. The etc. part is lat./lon., map references and foot notes for the highlighted objects in the field of view (FOV)

Human figures were obviously human and faces, while not regonizable as an individual, were recognizable as human faces. A pistol on a truck fender was recognizable as a US-M1911A1 model.

I suppose these could have been taken by aircraft but doubt it. Those were stamped "APRC, G-2 sec.", etc. as they were from Air Photo Reconesence Command, Group 2 section. In those photos faces were recogizable to the individual.

I can only assume that the technology has improved since 1974. So if you don't want to be photographed taking a pee, be sure you have a roof and walls blocking the satalites' view. :wink:

[joema]

Here are some KH-11 images. The 1st three were of a Russian shipyard, taken in 1984. Resolution is about 1 foot.

The 1st three images were leaked by Samuel Loring Morison, a Navy intelligence analyst. For this he was charged with espionage and imprisoned.

http://www.fas.org/irp/imint/kh11m_1.htm
http://www.fas.org/irp/imint/kh11m_2.htm
http://www.fas.org/irp/imint/kh11m_3.htm

http://www.fas.org/irp/imint/980820-O-0000X-005.jpg

[CJSF]

Human figures were obviously human and faces, while not regonizable as an individual, were recognizable as human faces. A pistol on a truck fender was recognizable as a US-M1911A1 model.

This sounds mutually exclusive. Is a US-M1911A1 pistol really so unique looking among pistols that you could tell what it was and NOT be able to see specific facial features?

[joema]

The problem with alleged facial recognition is it requires an oblique angle (unless subject is on their back). As the previously-posted images show, recon sats are capable of oblique angles, but that increases the distance (which hurts resolution), and incurs more atmospheric distortion.

E.g. the most credible perigee figure for the KH-11 is 273 km. At a 45 degree angle, simple geometry says the slant range is 386 km. At that range the diffraction-limited linear resolution of 2.5 meter optics at 500 nanometers is 9.3 cm (3.6 inches). That assumes 100% perfect adaptive optics AT VISIBLE WAVELENGTHS, which is questionable.

It's very unlikely you could identify a person's face (or a hand gun) with approx. 4 inch resolution. You might tell if they had a beard or not.

[Doodler]

The other thing to keep in mind about reconsats is that they don't operate in a vacuum (sorta). They are given priority and, in theory, backed up by human intelligence on the ground. You don't go snooping for someone starting with a spysat, you have someone on the ground give you general information about the target, then the NRO adjusts the spysat's orbit to put your target in the footprint of the bird. To get the ultra-high resolution, you're sacrificing the amount of area covered. At the highest resolution, you may only have a couple dozen square feet in view (someone feel free to calculate that, if you wish). You're not scanning cities at three point whatever centimeter resolution hoping your target comes into view, you're using the satellite to verify something you believe to be there or you're ferreting out a particular detail that just cannot be seeing from below, like missiles on a boat, or a shipyard with a new sub and some fancy extras strapped to the hull and whatnot.

Because of their fast orbits and narrow viewing apertures, spysats make terrible trackers. They are MUCH better suited to determining what's at a given location and a general view of any activity in the area in excruciating detail or telling you how badly you've damaged it after an attack. They are meant to monitor fixed locations, not individual people.

[Sheki]

Thanks for the replies everyone. I would have to agree with the analysis (basically that it wouldn't make much sense to do what I proposed in my original post). However, saying that it doesn't make any sense to do something is quite different than saying that something cannot be done. Previously I was under the impression that it was pretty much impossible for a spysat to achieve the resolution we are talking about. Now, I am merely predisposed to believe it is unlikely that anyone would bother.

Sheki