That has to be a myth. First of all, you can (geostationary orbit).
No. Satellites in geostationary orbit cannot be used for optical surveillance. They are 36,000 km away from the earths' surface.
The angular size of an object at a distance is calculated by the formula
d / D x 206265 = R
where
d is the actual size of an object in meters
D distance to that object in meters
R is the resulting angular size in arc-seconds
So a 1m object at a distance of 36,000 km is
1/36,000,000 x 206265 = 0.0057 arc-seconds
To put that into perspective, the Hubble Space Telescope (which is equipped with a 2.4m mirror) has a resolution of 0.1 arc-seconds (I have posted the formula and calculations at the bottom of the post)
Hubble is 17.5 times too small to resolve a 1m object from geostationary distance. If it were possible to use Hubble at geostationary distance, its best resolution would show objects smaller than 18m as a single pixel! To put it another way, for a geostationary optical surveillance satellite to resolve a 1m object at ground level, its optical element (mirror or lens) would have to be 42 metres in diameter!!
Also the trick with spy satellites was (and is) to have them in a low, highly declined orbit, so they will sweep as much of the planet as possible, really fast.
Unfortunately the downside is that satellites have to follow orbits, and those orbits are precise, predictable, and cannot be kept secret. This means that the countries you are spying on know precisely when those satellites will be over the installation you are trying to photograph right down to the hours, minutes and seconds, so they can act accordingly. This is what Lockheed Skunk Works at Area 51 knew when they were developing the SR-71/YF-12A (then known as OXCART) while the Soviets were spying on them. They knew not to have the prototypes and testing models outside at certain times of the day. (Even so, they still boobed in a rather amusing way
https://news.nationalgeographic.com...ecret-hid-craft-base-declassified-a-12-plane/) *scroll down to "Shadows of Area 51"
However, the problem at the time of the cold war was the quality of pictures. Transmitting bandwidth from satellites was still rather limited, and the resolution of electronic cameras was low, actually deplorably by our present standards. So it was still used to take analog pictures on film, then have the satellite develop them and either drop them back down on Earth or transmit them by a telefax-like slow scanning method.
- So in certain instances, the U2 might still have been preferable. Especially when not over central enemy territory, so the risk of getting shot down was lower.
It doesn't matter how advanced digital imagery gets, the Laws of Physics has limits, and one of those limits is the wavelength of visible light - it requires "X" amount of optical element diameter to get "Y" amount of resolution and there is no way around it.
* * * * * *
Formula and calculation
The theoretical resolution of a telescope is calculated using the formula
R = 11.6 / D
where R is the the angular size of the object in arc-seconds and D is the diameter of the mirror in centimetres. The HST mirror is 2.4 meters (240 cm), so we can calculate that its theoretical resolution is 11.6 / 240 = 0.05 arc-seconds. (For comparison, the diameter of the moon as viewed from the Earth is half a degree, about 1800 arc-seconds
There is a slight problem with this however. Due to factors involving interference patterns and the wavelength range of visible light, the smallest resolvable object is about twice the theoretical resolution. This is given using something called Nyquist's Theorem; you can read about it here.
http://en.wikipedia.org/wiki/Optical_resolution#Sensor_resolution_.28spatial.29
So effectively, the HST's resolution is about 0.1 of an arc-second.