[Wavytone]

Conceptually it would be possible to image the entire night sky, every night, using an army of robotic internetted smallish scopes. I'd make them solar powered, with a panel to charge and a battery for storage, and a wireless data connection to whatever carrier is available. Programmed to take a bunch of images suitable for stacking around local midnight and dump the images onto a server in the cloud.

Assemble enough of these scopes to provide full coverage of the sphere formed by the night sky, allowing for some overlap. This is a compromise in therms of focal length and sensor size vs the number of scopes, but definitely within the realm of doable. If each camera covers 4 square degrees it would require tad over 10,000 scopes. Programming each scope to take images of say ten portions of the sky instead of 1 per night reduces the number of scopes required by a factor of 10, say 1,000. For each chunk of sky to be imaged, lets say 10 shots at 30 seconds apiece, so each camera would take 100 shots each night, for an imaging run lasting about an hour.

Now the locations. Assuming we could install them at any point on the globe, have every scope imaging at their local zenith. That way the entire dark side sky would be covered once, every night.

However there is an issue with inaccessible or insecure locations (oceans, the poles, Himalayas and deserts etc) and locations that are hopelessly cloudy. For the cameras that would have been located in these, merely relocate them to a more suitable location where there is internet access - but still have them aimed at their designated patch of sky by hour angle and declination.

Once deployed, at the server side, stack the images from each camera for each night. The result is an image of the entire night sky, every night. With 1000 cameras producing 100 frames per night, that's 100,000 images per night requiring a potentially a few terabytes of storage per day. A petabyte per year. And the backend has to munch through this lot faster and dump the old frames than the images come in - or it will be inundated.

Now... as for what to do with that data... deploy a web application in the cloud (utilise the massive computing power in the cloud) which takes each group of 10 frames, stacks them to produce the final image for each chunk of sky and compares the images night to night for changes, eg supernovae, asteroids on collision courses with Earth, alien spacecraft... whatever.

Sounds like a plan - anyone with a server farm lying idle, and friends in high places in Google, or AWS ?