Cold War Satellite Pics Declassified

wwwssabbsdotcom writes "Looks like 25 years ago, we were taking pretty good B&W pics of the rest of the world, interesting story. How about those Cuban Missile Crisis pics, do they have that roll available?"

Comments

kodak instant moments?

[greechneb]

According to the article they were looking for "Kodak Instant Moments" - I wonder how they would use that in a commercial. "When want the best images of your enemies, use kodak film..." naah.

Whose looking in your window?

[Havoc'ing]

If you figure we can view a galaxy a bizzillion miles away through the hubble just imagine what we are capable of now right in our back yard. And the hubble aint even classified.

Re:Whose looking in your window?

[phil reed]

just imagine what we are capable of now right in our back yard.

Not as much as you might imagine. A Hubble-sized telescope in orbit at Hubble's altitude, pointed straight down, can resolve down to 15 centimeters. That would be enough to tell that you drive a Honda instead of a Surburban, but it couldn't tell much beyond that.

Re:Whose looking in your window?

[phil reed]

Well, my comment was based on the laws of physics. You can throw as much technology as you want at the problem, but it's physically impossible for a Hubble-sized mirror, looking straight down from Hubble's altitude, to read a newspaper headline. You would have a hard time even telling that you were looking at a newspaper.

We should avoid using spy movies as a basis for estimations on what our government is capable of.

Re:Whose looking in your window?

[phil reed]

Here's an older DejaView message on the topic. It addresses this very issue.

Undergraduate physics:

Resolving power R (resolution) of a diffraction limited telescope: R = wavelength/(2*diameter telescope)

This means for the HST (2.4 meter) and visual wavelenght (500nm) R = 500nm/4.8m = 1*10^(-7)

Since the Hubble is in orbit h = 680km (380 miles) high, this means it can theoretically resolve: Detail = R * h = 0.07. Thus 7cm (3 inch) details. Not enuff for reading license plates, even if someone would hold it up to the sky so we dont have inclination effects. Besides this, the best visual wavelength camera on board (the PC chip on the WFPC2 camera) UNDERsamples this signal by a factor of 2 giving an effective resolution of 14cm (1/2 feet).

This holds only if we ignore atmospheric turbulence effects (which certainly DONT average out), the pointing instability (up to about 10mas (micro-arcseconds)) and thermal breathing (up to 10mas). This degrades the image even further. (10mas translates to about 5cm as seen from the HST)

Furthermore, target acquisition is problematic. HST uses guide-stars, which need to be in the field of view, to lock on targets. Certainly no stars available on the face of the earth ;-p. Even then: A quote from the HST data hanbook: "It is also possible to take observations (primarily WFPC2 "snapshot" exposures) without guide stars, using only gyro pointing control. The absolute pointing accuracy using gyros is about 14" (one sigma), and the pointing drifts at a rate of 1.4 +/- 0.7 mas s**-1. "

So, we have a 66% chance of 14" (arcseconds) acquisition accuracy. This translates to about 1400 pixels offset (if we were well sampled) on a ccd camera or 100m inaccuracy on the ground.

Say we want a spy satellite with 1cm resolution (ignoring degrading affects) on orbit 300km high (if lower, atmospheric friction would cause it to fall back to the earth), then applying the same formulae as above we would need a telescope diameter of roughly 5 meters. (According the the space shuttle reference guide you would have to keep the payload doors open in flight to make it fit, hehehe)

Conclusion:

IMBO the NSA cant read license plates. The technology for space telescopes with this capability is only now being developed (look for NGST on the web) against HUGE costs, certainly not within the NSA's budget. Besides target acquisition is a severe limitation, and it's role becomes more important when the resolution increases.

just my $0.02

Re:Whose looking in your window?

[mesocyclone]

Problems with this analysis:

1) Id doesn't take into account any electronic processing of resulting signals or using multiple images taken seconds apart to achieve higher resolution.

2) Optical spy satellites are likely to use multiple mirrors, both to use adaptive optics to adjust for atmospheric turbulence, and to avoid the problem of fit in the shuttle payload.

I believe it is the Keck telescope (which has adaptive optics) that has resolution sufficient to read a license plate from much higher orbits (all other caveats apply).

A little tidbit... the Multiple-Mirror Telescope (MMT) on Mt. Hopkins, Arizona used to be (and may still be) owned by the airport. When it was built, it was built with an azimuth/elevation mount, rather than the usual polar mount, and used a computer steering system that was accurate enough to account for minute flexing in the very rigid metal frame.

The mirrors were Air-Force surplus from the spy satellite program.

The Air-Force used to "borrow" the scope from time to time. The Az-El mount was probably chosen to allow tracking of earth orbiting objects - Russian satellites.

The spook folks work with very impressive technology. They are bound by the laws of physics, but they probably have engineering tricks that the public world has not heard of. Tricks in signal processing, adaptive optics in space, ultra-precision pointing, etc.

Actually, if you just solve the problem of taking current earth borne adaptive optics telescopes into orbit, you can pretty well achieve the resolution you want.

And then, of course, there is synthetic aperture radar. Synthetic aperture is a mathematical technique for creating a synthetic (virtual) antenna of very long length (very high resolution in one dimension) along the motion of the radar. Simple radar has, of course, much lower resolution than optics for the same size antenna, due to the much longer wavelength. But when you extend the antenna for hundreds or thousands of meters through synthetic aperture magic, that resolution gets very good.

And then, of course, we can speculate about Lidar. I have no idea what the spooks may do with that.

I think the problem of resolution is no longer of significance to the spook business. The bigger problems are areal coverage, data reduction and storage, and concealment.

Might as well say it....

[LittleGuy]

"I can see my house from here!!!"

Oh good!!!

[Pig Hogger]

Will we finally be able to see Jackie Kennedy's pix while she was sunbathing on Onassis's yachts???

ho hum....

[Anonymous Coward]

*Mozilla 1.0+, Netscape 6.0, and Netscape 7.0 uses the JAVA 2 Virtual Machine which does not support the original Java applet security model used with EarthExplorer applets.
**EarthExplorer will not currently work with Macintosh systems due to the following:

1. IE and the Microsoft Virtual Machine does not support LiveConnect for Macintosh systems.
2. Old versions of the Java Virtual Machine (Netscape 4.6 and earlier) do not support LiveConnect for Macintosh systems.
3. The Java 2 Virtual Machine does not support the original Java applet security model EarthExplorer uses.
4. Signed secure applets don't communicate properly through LiveConnect when using the JAVA 2 Virtual Machine on a Macintosh. See http://bugzilla.mozilla.org/show_bug.cgi?id=160274 for more information.

Macintosh users may search for many of the same products at: http://edc.usgs.gov/webglis. You can also access EarthExplorer using the PC emulation package "Virtual PC" if you have this installed on your system

Interesting, somewhat related

[Anonymous Coward]

While working at nasa, a co-worker told me this story once:
Apparently in the 80's, he had been working on a satellite which contained a sensor to measure ground temperatures. The contractors who were working on the image processing for the data were so far behind, that the program would not be ready until a couple months after the satellite launch (a major PR disaster - no pretty pictures for the public to see!). So he was put on a crack team to hack something together that would be ready by launch time. What they ended up putting together was better than the specs. So the satellite launched and they got back the pictures and saw alot of interesting things... Like, gee, what's that underground hot spot in Nevada, and so on and so forth... So they were all pleased with themselves until the Feds came, classified their program and all the images, dumped all their equipment in a truck, and drove off.. I guess this shows why it is never better to do more than "government work" :)

Other News

[stinkydog]

In other news the website http://earthexplorer.usgs.gov/ has been crashed by unknown terriosts from the shadowy 'slashdot' organization. This massive 'Denial of Service' attack, know as the 'Slashdot Effect', is the orginizations trademark, much feared by webmasters and network engineers everywhere.

SD

Re:Cuban Pics

[meringuoid]

Yes, the intelligence pics that proved the Soviets had missiles in Cuba were taken by U2 spy planes. They were published immediately - if you're trying to force the Russians to remove their missiles, you don't keep it a secret that you know about the missiles. You tell the world.

In the early sixties, satellite reconaissance was primitive - it was still at the stage of ejecting the film in a little capsule to be picked up on the ground :-) Planes were getting much better material then.

The probable limit of spysats

[MtViewGuy]

I think the practical limit for today's KH-11 and newer spysats is about 6-7 cm resolution, not enough to read a newspaper headline but definitely good enough to tell what kind of vehicle you're looking at.

Remember, even at 100 cm resolution the IKONOS satellite is capable of showing some amazing images. Remember that IKONOS image of the North Korean rocket test facility?

I expect within the next 4-5 years several companies will be orbiting imaging satellites capable of resolution at 100 cm resolution. It'll be nearly impossible to hid any secret activity with that type of resolution.

KH-9, Big Bird

[wiredog]

It was the last of the bucket droppers. An interesting book, if you can find it, is Deep Black. It's a history of overhead imaging from the Civil War through the KH-11 program, including the U-2 and SR-71 aircraft.

My father worked for the Defense Mapping Agency (the predecessor of NIMA) until 89 and he was surprised at some of the things that showed up in that book. Especially that the resolution of the KH-11 (best is 2.5 inches, so it can't read license plates) and KH-9 (9 inches) were in there.

USGS web page: Gale Norton strikes again!

[mfago]

As mentioned in another post, the USGS webpage itself is unusable unless you're running Netscape 4 (windows or linux only) or IE for Windows.

I think it would be a good idea for as many people as possible to emailthe maintainer of the web page.

Unsurprising for the gov't to so thouroughly screw-up like this, especially with Interior Secretary Gale Norton at the helm. FWIW, she is facing contempt of court charges for lying in Federal court during a trial of gross mismanagement of the Native American Trust fund. Mismanagement by completely failing to secure a computer system...

Hell, why don't we all email Gale herself?!

Yup, the film was Kodak film -- no kidding

[Buran]

Actually, the Corona program did use Kodak film. Due to static problems with early film (which caused arcing on the exposed negatives), Kodak developed polymer-based film.

I work in an electron microscopy lab and the film used for the EM systems is Kodak 4489 "ESTAR Thick Base" -- which means that my paychecks depend directly on something that was developed for use in space. (As a space buff -- Buran is/was the Soviet space shuttle -- I'm quite pleased with that situation.) A spinoff, as they're commonly called.

The EM film is mounted on metal plates for exposing and when developed yields 8cmx10cm transparencies using Kodak D-19 developer. For Corona, the exposed film was placed in a reentry capsule which parachuted back to earth and was retrieved midair by a C-119 Flying Boxcar aircraft. It doesn't take that long to develop at all and can be ready for analysis the same day.

According to the Kodak EM film page:

"KODAK Electron Micrography Film 4489 has approximately half the speed of KODAK Electron Image Film SO-163 film, but exhibits less curl and shorter pump-down times. Coated on a 7mils Estar support, KODAK Electron Microscope Film offers exceptional dimensional stability and eliminates the use of traditional glass support products."

We are still using film because (1) electron microscopes are very expensive, so ours are from the mid-1970s, (2) it's not that easy to retrofit them, at least as far as I understand it, for full digital, and (3) it's not all that hard to put the negatives on a lightbox and shoot them with a professional digital SLR, which is how we get the images into computers for processing. And, of course, (4) digital camera technology still hasn't beat out film for quality yet, though we're hoping to get a Canon EOS-1Ds soon that will start to close the quality gap.

(The film is kept in a vacuum once in the microscope -- something else which I'm sure was a benefit for Corona.)

If you want to see some sample EM images taken with the Kodak film, see our lab's image gallery. Don't bother with Kodak's sample images, they suck. ;)

I'm pretty sure that Kodak also designed the Corona camera system, though I'm not certain who the actual builder was.

Panoramic Imagery

[briancnorton]

This is all great and all, but I have worked with corona imagery (after 1996) and it's REALLY hard to use.

First of all, the imagery is not vertical, it's panoramic. Great for intel agencies, not so great for mapping. It's almost impossible to orthographically rectify, and hence use for anything useful. The resolution of the film is very good. It's something like 150 lp/mm, and the stereo is very good, but it's a pain in the butt to do panoramic stereo without special equipment.

second, geo-referencing was accomplished in a brilliant, if arcane way. A second camera was involved that took pictures of the stars 180 away from the image. To find out what the picture is of, you need starcharts and a lot of math to figure out what stars you are looking at, where the satellite was, and what the picture is of. The equipment to do this in a useful environment is VERY expensive.

third, it's panchromatic and not IR sensitive. You can see some ground features, but nothing environmental, and not all that much of historical significance. Consequently, the imagery has not been used for as much as had been hoped.