Whenever I watch these tv shows (24, for example) they always show them using satellites to spy on people. They can see faces, read tags on cars, and other stuff. Many people claim that this is accurate. Yet, when we put an orbiter around Mars, something the size of a car is hardly bigger than a dot. Why? Mars has hardly any atmosphere to cause blurring. Its a smaller planet, with less atmosphere so I assume the satellite can even orbit at a lower altitude. Does NASA get the cheap stuff or are we being lied to about how good spy satellites really are?

Lied to by tv shows.

When you look at the stars you see twinkling due to the atmosphere.
The atmosphere causes the same blurring when the satellite looks down, there’s a limit to the useful resolution.

I’m reading the 50cm per pixel is considered good for a satellite.

Note that Google maps, google earth and other map systems say its “satellite” but it also contains photos taken from airplanes…
And neither of them and no website is “real time”…
And tv also lies about de-pixelisation. You can smooth the edges off the pixels, but you can’t turn pixels into true high resolution, it would still just be low resolution … just blurry looking, with less obvious “artifacts”’ (artifacts are false lines, and false features … such as herring bone patterns and spots and so on.)

Even with no atmosphere, satellite (and all other) image resolution is limited by diffraction. That gives angular resolution on the order of lamba/d radians, where lambda is the wavelength of light, and d is the diameter of the telescope. Earth-orbiting spy satellites have bigger telescopes than Mars probes, because (a) the military has more money to spend, and (b) they don’t need to send their telescope as far.

Even the military almost certainly can’t read license plates in any useful sense. The details are classified, so we can’t know for sure e.g. what clever image processing they have, but I doubt it.

NASA’s eigenvalue is off?
An excuse to post a great compilation…

As a rule, governments do not want anyone to know the extent of their capabilities. 'tis a secret!

Although it is possible with certain imaging techniques to get better than 50 cm/pixel resolution, there is no way to get a resolution at MEO (where most surveillance satellites are located) to read the text off of a license plate (not withstanding that license plates are generally oriented close to orthogonal to the optimal viewing angle for visual surveillance satellites. Atmospheric distortion also places limits on the best possible resolution, albeit not specifically because of distortion due to turbulence (which causes the twinkling of the stars) which can be largely compensated for by a combination of active (adaptive) optics and wavelet processing (which can also be used to depixelize signals with regular patterns); the limiting factor is differential diffusion and absorption across the spectrum of visible light.

In theory, satellites in LEO should have better imaging capability by virtue of being close to the illumination source; however, most LEO satellites have much smaller light gathering capacity (smaller optics) which limits their effective resolution; however, the next generation of high geometric fidelity inflatable optics may rival fixed telescopes and offer essentially unlimited light gathering capability, so the limit will be atmospheric diffusion/absorption.

Satellites generally speaking are not used for real time visual imaging the way you see in action-thriller movies, although they are sometimes used for IR and radio frequency tracking. The Defense Support Program (DSP) satellites that are used for launch detection and early warning launch surveillance look for characteristic IR signatures associated with rocket plumes.


So spy satellites can’t read licence plates and never will (could they do active compensation of atmospheric distortion?), but we’re missing the other part of the OP. Is the orbiter equipment the best we could do, or have other constraints led to choosing imaging technology that’s “good enough” for the pictures NASA wants of the surface?

The best images of Mars come from the HiRISE instrument on the Mars Reconnaissance Orbiter. It’s 0.5-meter aperture telescope. (A telescope’s resolution is proportional to the aperture, unless limited by atmospheric distortion or build quality.) It weighs only 65 kg, which is incredibly light for a telescope that size.

We don’t know how big the current spy satellites are, but it’s widely believed that the Hubble Space Telescope is based on the KH-11 spy satellite series. The Hubble is a 2.4m telescope, so it should be almost 5 times better resolution than the HiRISE. The Hubble weighs 11,000 kg.

This. American KH series of spy satellites are similar to the Hubble in size and weight (they’re just pointed back at Earth rather than at the stars). Putting something that heavy into low Earth orbit still isn’t cheap, but sending one to Mars is simply not possible.

I would say it is a little better understood than “widely believed” now. Given NASA has been gifted a pair of KH-11s by the NRO, the precise capabilities and differences are understood. There is a lot of commonality. 2.4m diameter mirror being the obvious one. But also a lot of differences, reflecting their different roles.

[quote=“Riemann, post:4, topic:755437”]

NASA’s eigenvalue is off?
An excuse to post a great compilation…


I always preferred this one.

50 cm resolution used to be the limit that the US government let civilian operators give out, but they recently lowered it to 25 cm due to competition from foreign operators (I don’t think any civilians have 25 cm resolution yet, but it’s getting better). WorldView-3 has a resolution down to 31 cm, with a 1.1 meter mirror. Modern US spy satellites are thought to have a 2.4 m mirror. It’s thought the maximum resolution is around 15 cm, give or take. Enough to see what color shirt you’re wearing, but not to read the logo on the front of your shirt (or your license plate number).