Hacker News new | past | comments | ask | show | jobs | submit login
Ask HN: Which is the most powerful satellite camera directed towards earth?
48 points by o2l on April 18, 2017 | hide | past | favorite | 44 comments
Remember the scene from MIB where 'K' zooms in on his wife and watches her live. How close are we to that ?

After a quick search, I found a few videos which showed scenes of an area with some traffic activity but it wasn't that zoomed in. As far as images go, I haven't found any other image source providing a better definition image than what Google Maps provide.

The American government had used live satellite feed while carrying out the Bin Laden operation ( Source of Information: Documentary | Correct me if wrong ).

Are there any images/videos publicly available which demonstrate the limits of cameras in space directed towards earth ?

If not publicly available, can you describe how a possible image/video might look like and what are the current limits for those cameras ?

I worked a earth imaging system (EIS) back in the 1990s. I didn't have security clearance (almost got it and realized it had no benefit to me but a bunch of crap I'd have to do, or not do/say, so I passed).

I'm pretty sure that the satellites at the time could read 3 inch newspaper print under the right conditions which happened rarely.

I've got a tack sharp 600mm Canon lens and it's sort of useless for a lot of stuff. If there is any haze, heat, dust, whatever, in the air, all that expense glass sees is that instead of what you want to see.

That note aside, it was a fun project. I was the I/O guy, I did this work:


and I got Seagate to redesign part of their 9gb (I think, might have been 2 or 4g) barracudas. I've got a benchmark in lmbench that shows how the disk performs as a function of seek distance, looks like this:


The lower edge of the band is what you get if you seek and get to the track just as the sector you want is to about to pass under the head; the upper edge of the band is what you get if the sector had just passed under the head; the height of the band is a rotational delay; and those outliers? In this case I think they were either bad blocks or I don't know.

But when you ran this benchmark on two drives, mounted in a rack right next to each other, you got tons and tons of outliers which blew any chance we had of meeting the performance metrics. I bitched at Seagate and they hemmed and hawed and finally admitted there might be a problem with their internal mounts.

The problem was that their mounts were so useless that the vibration caused by one head moving vibrated the drive next to it enough that the other drive's head didn't settle properly and you blew a rev waiting for it to get where it needed to be.

Seagate redid the mounts.

Fun project.

I remeber flipping through some spy tech book a while ago and one photo, allegedly quite famous, stuck with me - it was of a wrist watch on a hand of someone lying on a beach and while it was on a blurry side you could still tell what the time was. From the context it was taken in the 70s and from what I gathered it was a satellite snap.

Now thinking about it, it might've been a high-altitude plane (U2) photo instead, but I'm not sure.

Has anyone else seen this photo?

I seen it, but it is a famous fake. A few centimeters is the lowest bound for resolution because of atmospheric distortion.

Also looking right from the top, at the best conditions, from 200km (a stable orbit can't be lower), will give 4cm resolution for a Hubble-like mirror, or 2cm per pixel Niquist frequency.

That really rather sounds like from within the atmosphere - heat shimmer alone should make this sort of resolution very difficult indeed from space.

>> I bitched at Seagate and they hemmed and hawed and finally admitted there might be a problem with their internal mounts.

And now they are happy because RAID configurations are fairly common these days.

Satellites have had the ability to do that for decades. I was in the military during the Reagan administration and we would roll $weaponsystem under cover when the Russian satellites were passing over. They could see even small details such as which model it was.

But more resolution isn't what's important. Other capabilities, such as deep infrared, ultraviolet, radio, whatever scanning is more useful. Sats made U-2s, SR-71s mostly obsolete; drones are doing a lot of what Sats did. Top of the list though is refueling spy sats. That is the Holy Grail of reconnaisance which the USAF X-37B totally experimental research craft totally doesn't do.

Refueling massively extends their use and lifespan. Its like have twice as many for the money. Cuz, ya know, a hundred ain't enough.

There are some current projects out there where a vehicle is launched that can attach to most satellites. It has it's own fuel to do station-keeping. I know there are some other projects involving a vehicle actually refueling satellites and then deorbiting but I'm sure it is not compatible with everything out there.

what does "refueling spy sat" mean? (I know nothing about satellite power sources)

The problem with spy sats is they orbit too low, so the suffer from drag in the trace atmosphere. Also, moving them in their orbit is super useful. Covers more, plus confusing the opposition as to their location. But it takes too much fuel. If you refuel them, both problems solved.

They have onboard fuel for "stationkeeping"; most low orbit sats do. That is keeping them in their proper orbit and attitude in spite of drag.

I would expect a more elliptical orbit to be useful too. It would let you swing down lower over your target while keeping your satellite out of the atmosphere most of the time, and it would be even harder to track.

The trouble with using elliptical orbits to get low and close is that you're going incredibly fast at that altitude. This spells trouble both because you spend a small fraction of your time above your target, and because drag is proportional to the square of velocity, mitigating any savings there.

We do have uses for highly elliptical orbits, but they're usually used for long residency times on their high side. The Molniya orbit is a prototypical example.


Tracking is not difficult for groups of amateurs with telescopes. Considering that you are up against large foreign states with extensive radar and optical detection facilities, it is mostly a lost cause. Satellites have enormous solar panels.

Elliptical orbits still require station keeping. Elliptical orbits tend to circularize over time from the drag they experience at the closest point to the Earth.

41cm/pixel available to the government & 50cm/pixel available to Google through GeoEye-1


31cm/pixel through WorldView-4


take it for what it's worth, but according to this bbc article published in 2014...

> "The latest US spy satellites, in comparison, are reported to be able to pick out objects less than 10cm (4 inches) across."


Wow - I figured they were ahead, but 3x...

Some links for you:




A perfect 2.4-meter mirror observing in the visual (500 nm) would have a diffraction limited resolution of around 0.05 arcsec, which from an orbital altitude of 250 km would correspond to a ground sample distance of 0.05 m. Operational resolution should be worse due to effects of the atmospheric turbulence.


I'm assuming this is all pretty well classified, but I would be shocked if the US government hasn't had the ability to do that for a long time. Consider that back in 2012, the National Reconnaissance Office donated two surplus telescopes to NASA that were substantially better than the Hubble telescope. If those are the sort of telescopes that the government has no need of anymore, imagine what the state-of-the-art is!

As an aside, that donation has actually put NASA in a bit of a bind. For political reasons they can't very well turn down the offer. But the telescope itself is only about one third of the total cost of space telescope, with the rest being due to the cost of the instruments and the launch itself. Unlike the US military, NASA does not have an unlimited budget so this unexpected expense threw a wrench into their long-term plans. Furthermore, the telescopes were designed to look down instead of up, so they're not optimized for astronomical observations.

Not sure why you're being downvoted for this. https://en.wikipedia.org/wiki/2012_National_Reconnaissance_O...

Hubble was launched in 1990, is it surprising that 22 years later, NRO telescopes are substantially better ?

The donated telescopes were build for the Future Imagery Architecture project:

They're substantially newer technology than Hubble (which is older than the 1990 launch date suggests -- it was delayed several years after the Challenger disaster).

That said, I don't think they'll give an appreciably better resolution -- given good fabrication techniques, optically a 2.4m mirror is a 2.4m mirror, and everything I've read suggests that US IMINT satellites have been (in good seeing conditions) close to diffraction-limited for a long time now.

I'm sure the capabilities of the satellites have improved a great deal, but in other directions than resolution. More communications bandwidth? Faster repointing (giving more flexibility about which targets get imaged on a given pass? Better multi-spectral imaging? I'm sure there will be some surprises when the program is eventually declassified. (And I'm optimistic it will be -- there's lots you can read now about the film-return satellites).

The point is they were giving them away. Meaning the sats they kept are even better than better than bubble.

The satellites would've been near contemporaries of Hubble, not 22 years junior. NRO didn't need them anymore.

There is a question on the Stack Exchange network which asks about publicly-viewable satellite cameras: https://space.stackexchange.com/questions/4485/which-high-re...

Disclaimer: I'm the OP on that question.

High-rez satellite cameras don’t film videos.

Satellite cameras don’t have a 2D sensor you can find in a traditional camera. Instead, they feature a long narrow sensor. As the satellite flies over the earth, the sensor records a single row below the satellite. That row is oriented perpendicular to the satellite’s velocity. As the satellite flies, it scans a long strip of land underneath.

It works similar to a flatbed scanner. Can you film a video with the scanner? I don’t think so.


About Bin Laden, the live satellite feed was not filmed from space. The feed was recorded by an on-body camera of an American soldier, transferred to a telecom satellite, then back to Earth, at Obama’s place.

Video from space has been done by civilian spacecraft, e.g.:

We know that the film-return satellites were working in push broom mode (or "whisk-broom", as I've seen the KH-9 cameras described). I have little doubt that the primary observation mode of the follow-on electro-optical systems is also push-broom. But not sure I'd want to rule out the possibility that the more recent blocks might also have some kind of staring mode (which could plausibly give video of small-ish areas). They're still pretty secretive birds.

The resolution is not that great, 1212 pixels / 2.5 km gives 2 meters / pixel.

The ground wind speed at the time of eruption was below 5 m/s, i.e. the video plays much faster then real-time.

If you’ll download the video, pause it, skip between frames, you’ll notice different parts were shot from different positions and with different resolutions.

They probably abused the fact clouds move in very predictable manner. They took several satellite images from whatever satellites happen to pass roughly above the volcano, and used creative video editing to make a smooth video from those several frames.

Sorry, my bad. They used drones for that.


This article explains more about how they used a combination of satellite imagery and drones.

Yep, unlike satellites, modern UAVs can deliver a live video feed just fine.

I suspect the most powerful are classified NRO satellites. The public's knowledge is likely a decade or two behind actual capabilities.

I recall reading somewhere about satellites being able to see license plate numbers from space but I have no idea where I read that or if it's true.

I'm pretty sure it was NRO that was our customer. Any idea if around the late 1990's the NRO was run by a woman? If so, that's the agency we were working for (as mentioned in another comment).

KH-11 had a theoretical resolution in the inches. "Not quite good enough to identify faces." Those were first launched in the seventies, so I assume they are doing much better now. I think it's safe to assume 5cm or better resolution on a minutely refresh.

makes you wonder why the imagery floating around with reporting on war zones and incidents are of such bullshit quality.

even the supposedly recorded-with-mobile footage is always so bad. as if the person is stumbling around not knowing their phone's cammera is rolling.

I assume militaries don't provide news outlets with their highest resolution spy imagery as to not disclose the extent of their capabilities.

I'd also suspect anyone recording a news-worthy "incident" on their mobile phone in a war zone has other priorities than maintaining a stable image.

> makes you wonder why the imagery floating around with reporting on war zones and incidents are of such bullshit quality.

Hard to say what you're talking about without examples, but stuff like military targeting cameras are frequently 1970s tech, and looking via infrared to boot. There's plenty of high-res footage from war zones available.

> even the supposedly recorded-with-mobile footage is always so bad. as if the person is stumbling around not knowing their phone's cammera is rolling.

Being shot at has that sort of effect. That said, there's some pretty remarkable 4k drone footage coming out of ISIS these days, along with lots of GoPro footage.

> As far as images go, I haven't found any other image source providing a better definition image than what Google Maps provide

As many have alluded to here, there's likely to be a big difference between publicly available imagery and what's currently possible. Sticking to publicly available stuff though, while the resolution of all mapping services varies a lot across the globe, http://wego.here.com provides satellite imagery that betters Google's in a lot of cases.

This is all somewhat moot if you're looking for MIB-esque video of course as these still images are compiled rather than directly snapshotted

It's hard to compare Here Maps and Google Maps here. It's out of question that Google Maps provide a superior experience as compared to Here maps. But all the Google Maps imagery is processed, even while using the 2D option ( This is true only for countries with street view, from my observation. It's brilliant how they use street view along with satellite imagery to make a 3D model ). For my country India, the quality of images are slightly better on Google Maps, whereas when I check for other locations like Beijing, Here Maps seem slightly better than Google. It's weird how image quality change from country to country. It must have something to do with how they use their resources to provide best quality to probable customers and also country specific laws might apply for releasing satellite imagery.

I expect drones are now more relevant than satellites for localised imaging.

For example http://www.cropcopter.co/uav-imagery-vs-satellite/

Yes. There was a piece on 60 Minutes a few years back where they had a big (~20 ft. wingspan maybe), loud drone circling overhead a few thousand feet up. It was inaudible and invisible. What's the point in solving a hard problem like hi-res imagery from space when you can just fly one of these over the target practically at will? (With a Hellfire missile attached, no less.) Range and airspace issues notwithstanding, the drone future is a scary place.

> Range and airspace issues notwithstanding

Those are precisely the reasons satellites will remain relevant. Satellites are the reason the SR-71 fell out of service. Russia can and will shoot down US aircraft or drones in their airspace.

SR-71 is about as far from a drone as you can possibly get. Politicians are going to (pretend to) care a lot less about airspace sovereignty when violating someone else's airspace does not involve risking pilots' lives. Range and stealth are only going to increase. This is why I said the future looks bleak.

The problem is the distance involved and the orbits. For example, GeoEye-1 is in a sun-synchronous orbit in order to be over the US at the same time each day. However, it means that the satellite is moving quite rapidly across the surface of the earth much of time. If you've taken a photo of a moving person or car, you know how hard it is to get a non-blurry picture. Now, try doing that at much higher speeds. If you put a satellite into a GEO orbit, it is incredibly far away from the Earth and would require huge optics in order to get good photographs.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact