This is necessarily akin to asking the length of a piece of string.
Still, anyone have a handle on the likely bit rates that are achieved when downloading data from low Earth orbit imaging satellites?
We know, or at least think we know, that the USA, and other countries, have spy satellites capable of significantly higher resolutions than those offered by commercial providers, but I wonder what scope they have to providing high resolution imagery over large areas (e.g. southern Indian ocean). I suspect that it is constrained by downlink bandwidth.
They say DVB-S2 is as close to the Shannon limit as anyone has achieved.
Photo interpretation is tougher than it looks. It isn't like you get to look at the debris from multiple angles, but rather you see a 2D image. If you go through the Google Earth forums, you can see all sorts of nonsense of what people think it is there versus what is seen by a person on the ground. I have seen people confuse craters for mounds (i.e. concave versus convex.)
Look at the specification for some of the recent TDRS satellite. The TDRS satellites are geosynchronous and receive data from low orbit satellites and sends the data back to fixed ground stations.
Not sure the geosynchronous satellites tell us that much.
The low Earth orbit satellites move across the sky (quite fast, indeed), so the higher the antenna gain (both on Earth and on the satellite), the more accurately it must be pointed at a moving target.
On the other hand, it's a lot closer, which helps with the signal strength.
At least the early TDRS satellites had multiple antennas and receivers for a specific band. The individual antenna signals were then moved with frequency translator to a big downlink.
On the ground station, individual signals from the different antennas were then extracted and frequency translated from the big downlink and combined with various delays to create an electrically steerable antenna lobe. Such electronic beam forming does not have problems tracking the LEO satellite, which is in the field of view for about 30 minutes.
Depending on the ground equipment capability, multiple satellites can be simultaneously tracked, even if they fly in the opposite direction i.e. N->S vs. S->N.
Once upon a time, NASA had tracking stations scattered around the world in the STADAN network tracking near earth satellites. In order to save money and give real time control of the birds, NASA launched the TDRSS network to relay the data direct to the US no matter where they are in orbit, and shut down the STADAN stations including Orroral Valley in the ACT.
Play some games: in late 60's could read a newspaper headline lying on the ground, later read a license plate numbers at an angle, and now public domain videos/images clearly show infrared signatures of 'victims' as they're dismissed, and assuming the resolution of those images is trashed to obfuscate true capability. leads to...
Assume with zoom the capability is smaller than 1/8 inch resolution, but over what field of view? Arbitrarily assume 1000ft by 1000ft field of view, why so wide? well, just how much would thermals 'wiggle' the image? Need to encompass the total image for the post processing to remove those pesky thermals. Also, early 70's it was pretty easy to obtain almost lossless compression ratios of 30:1, maybe algorithms have gotten better, assume 50:1, or even 100:1. How many frames a second? Again, assume 1 to
10 fps, that is 'update', not aperture time. Envision, flash, transmit, flash, transmite without blurred motion.
]digress for a bit: How does that compare to a 10M pixel over-the-counter camera? assume a field of view of 10ft by 10ft at distance of 10ft, yields
40 mil by 40 mil pixel. not bad. What would that be over field of view of
1000ft? approx 4 inhes by 4 inches, so resolution would have to be 30 times, or is that 900 times, better. Seems doable. ]
With all those 'assumptions' what would one get when zoom back to a wider field of view of 1000 miles by 1000 miles? That high resolution camera would then yield a resolution of 55 ft. and a 747 would subtend around 4 pixels by 4 pixels, kind of small, eh?
Ok, assume something interesting was happening over that 100 miles zone, then the resolution would be around 5.5 ft by 5.5 ft, and a 747 would subtend 36 pixels by 36 pixels. not bad, but still out of a total pixels of 10G pixels, might take a bit of smarts to find.
Back to BW required to send an image: assume a 100:1 compression, 3 colors at 32 bits per color, and 10G pixels, and update at one per second means the transfer rate would have to be approx 8.8 Gbps, not including any framing. Again, that's up there but achievable at 100GHz transmission rates using clever modulation schemes. Trade off update for BW, trade off resolution, or compression, for BW and might make it happen.
[Now, how big would the 747 look to our 'commercial' camera? 36 pixels by
36 out of our high resolution camera with 9G pixels compared to 10M pixels, would be like 36*sqrt(9000/10) or 1 pixel!!! That translates to ... take a picture of a blank wall and now find a single pixel that's not 'right' Again, doable, but a bit of a challenge.
That was fun. I put all the numbers in, in case made a mistake somewhere, and youse guys can check.
The optical resolution is ultimately defined by the aperture ratio (lense/mirror diameter vs wavelength). Unfortunately the atmospherics disturbances are going to kill the resolution far earlier.
This post reminds me of a young Australian entrepreneur I heard about on ABC radio recently who is implementing a "flock of dove" satellites (not much bigger than bread box) to map a continuous up to date image of the entire earth.
The satellites were prototyped with off the shelf smart phones, but the production models are more sophisticated able to be repositioned in orbit.
None of this makes sense unless you specify a bandwidth, power, and path loss. That is why I mentioned DVB-S2, i.e. how close you can get to the Shannon limit. That is, in any communications analysis, you need to to know the channel characteristics.
Atmospheric disturbances are insidious. However, the 'correction' algorithms used in post processing are AWESOME! Even in the 80's I saw pure crap turn into broadcast quality.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.