I know the pictures can be very good now, but how can they tell what is going on inside a building? I often wonder if they also have assets on the ground that help fill in the details but of course they cannot reveal that.
Back during the Cuba missile crisis the pictures were not that good but yet the US still knew what was going on.
They were mostly taken from reconnaissance aircraft (especially, the U2) and they were pretty good.
Did the cameras have x ray vision?
What makes you think they can?
As above, that was done with U2’s (one of which was shot at) and they were taking pictures of missile launch sites, which were not obscured inside buildings.
Slightly off topic, but somewhat relevant:
In the 1970’s I was involved in a project in which I had access to numerous satellite photographs of the then Soviet Union.
In one picture there was a clearing in the woods, and in that clearing a woman was lying on her back enjoying the sunshine, naked.
In the photographs, it was clearly obvious that she was a natural blonde.
During WW2 we had Operation Crossbow. They were taking photographs of Nazi installations ane rendering them in 3D.
You can’t directly tell what is happening inside a building.
But if you watched it being built or equipped you can know a lot about what’s inside. An office building and factory will have different fit-outs.
You can count cars & busses coming and going and establish about how many people work there. On which days and shifts.
You can observe shipments coming and going.
etc. Coupled with some HUMINT on the ground to calibrate all this you can learn a lot. Not everything, but still a useful incrememnt of information.
The optical resolution is very good, and was first publicly revealed in Jane’s Defense Weekly of August 1984. Normally, oblique look angles through the atmosphere would be greatly degraded, but apparently they use adaptive optics to compensate. These images do not represent the true capability but hint at what’s possible:
http://oi53.tinypic.com/141lcvt.jpg
http://oi53.tinypic.com/205b8t0.jpg
Although fictional depictions often present similar resolution in the infrared spectrum, two items in basic physics place severe limits on this.
(1) Atmospheric water vapor strongly absorbs IR except in a few windows around 3-5 microns and 8-12 microns:
(2) Those wavelengths are much longer than 390nm to 700nm visible light. For a given diameter optic, angular resolution is inversely proportional to wavelength, based on Dawes’ Limit. We can calculate the theoretical linear resolution of a KH-11 recon satellite (2.4 meter mirror) in 500nm visible light, 5 micron IR and 12 micron IR at an orbital altitude of 150 statute miles.
We first determine the angular resolution then convert that to linear resolution:
Angular resolution is from Dawes’ Limit: Dawes' limit - Wikipedia
Angular resolution formula:
a = 250000 x W / d, where:
a = angular resolution in arc seconds
W = light wavelength in meters
d = telescope diameter in meters
Visible Light angular resolution:
a = 250000 x 500E-9 / 2.4 meters (KH-11 mirror size)
a = .05208 arc seconds
5 micron IR angular resolution:
a = 250000 x 5000E-9 / 2.4
a = 0.5208 arc seconds
12 micron IR angular resolution:
a = 250000 x 12000E-9 / 2.4
a = 1.25 arc seconds
Linear resolution formula:
s = tan (a) x d, where:
s = linear resolution in units determined by d
a = angular resolution in degrees
d = distance to object
Linear resolution at 150 mi altitude using 2.4m mirror and 500 nm visible light:
s = tan (1/(3600/.05)) x 150 miles x 5280 feet per mile
s = 0.192 ft resolution using 500 nm visible light
Linear resolution at 150 mi altitude using 2.4m mirror and 5 micron IR:
s = tan (1/(3600/.52)) x 150 miles x 5280 feet per mile
s = 2 ft resolution using 5 micron IR
Linear resolution at 150 mi altitude using 2.4m mirror and 12 micron IR:
s = tan (1/(3600/1.25)) x 150 miles x 5280 feet per mile
s = 2 ft resolution using 5 micron IR
s = 4.8 ft resolution using 12 micron IR
So with IR, the image will be very blurry – even if the detector sensitivity and other sources of thermal noise permitted full use of this theoretical capability.
In recent years, with the advent of mobile devices, cell phones, cordless phones, microwave data links, etc. electronic surveillance is as important as optical. According to the book Body of Secrets by James Bamford, the Magnum/Orion class of electronic intelligence satellites use gigantic parabolic antennas 160 feet in diameter to pick up signals:
The previous calculations were just resolution, not considering the IR transparency of structural walls or roofs. In general I don’t think insulated outer walls or roofs are sufficiently transparent to IR such that ground, air or space-based thermographic imaging would be useful to see people, even as fuzzy blobs.
If they were outside they could be seen at night from space using either low light or IR sensors, but IR resolution would be degraded by the physics previously described. Even for this to work the atmosphere would have to be clear.
In narrow cases – say a person inside a sheet metal utility shed leaning against a wall, this can be seen via ground-based thermographic imaging. However if he was away from the wall he probably could not be seen from the ground, much less space. See attached.
https://joema.smugmug.com/Photography/Flir-Systems-thermographic/n-mnGn2j/i-HK5XqTJ/A
Also Francis Gary Powers in 1960
Police on the ground or in low-altitude helicopters have been known to detect marijuana growing operations from the high infrared output of a house. I see no reason in principle that you couldn’t do the same thing from a satellite. Though that’s not so much “seeing inside a building” as just noticing that something is unusual and making a reasonable inference. And of course, it’s really easy to shield against, if you’re so inclined.
So from this photo you could definitely tell that it was blond hair rather than shaved hair? I’ve gotta admit this is hard to believe for the 70’s. Making that distinction is harder than reading the headline on a newspaper.
Probably wasn’t shaved in the 70’s. 
I heard examples similar to that from a relative who was in a position to know, back in the 1980’s.
This depends when in the 1970s. The KH-11 was first used in 1976 and might under optimal atmospheric, lighting and orbital conditions had sufficient resolution. That satellite weighed about 36,000 lbs, of which 14,000 lbs was maneuvering propellant, so it had significant capability to adjust its orbit:
During the 1970s the previous KH-8 “Close Look” film-based satellites could use an Agena upper stage to temporarily drop their perigee to only 70 miles:
According to the November 1983 Popular Science, one was launched during the late 1970s Iranian hostage crisis which could identify specific people by the shape of their facial beards.
You can also tell a lot from the utilities going into a building. The electricity supply needed for an office building is quite different than that needed by a building full of nuclear diffusion machines, for example.
Exactly.
Overhead surveillance isn’t all knowing. But the whole point is to convert a large mass of total unknowns into a jigsaw puzzle where you have a bunch of pieces.
From there you can use logic & general knowledge to fit the pieces together at least provisionally. And just like solving any other puzzle, the more you know, the faster & more reliably you can know and integrate new information.
I very much doubt that those were satellite images: the laws of physics and optics limit the resolution way below the “reading the headline on a newspaper”/“identifying pubic hair colour” capabilities.
Reconaissance plane photos, I can believe, but satellites? No way, no how. The sheer distance and atmospheric thickness would prevent it.
Currently the best commercial satellite imagery has a resolution of about 40cm per pixel. I’m sure military satellites can go better than that, but not 10 times better.
The film-based Corona satellites back in the 1970s had a resolution of about 90cm: apparently 30cm was technically feasible but not practical. CORONA (satellite) - Wikipedia
Even the latest Keyhole satellites are unlikely to have effective resolving power below about 15cm. Maybe 10cm if you do some fancy deconvolving of multiple images.
Let’s have some pics of this blonde, please. It doesn’t even need to be that blonde. Just any blonde, taken from orbit. Not from a plane overflying a sensitive area when you’d expect them to be overflying the Ruskies. So far all we have is “I knew a guy whose cousin’s illigitimate daughter showed me these photos…”