Hacker News new | ask | show | jobs
by nradov 1682 days ago
Huh? I don't know why you think a convolutional neural network would help but that's not how this stuff works. Due to heating effects and clouds there is no "visual range" for a ballistic missile. Any terminal course corrections will be done with radar.
1 comments
sudosysgen 1682 days ago
That's incorrect. There is a visual range. You can either image in the UV spectrum, or simply choose your optics so you can image the carrier from before re-entry.

Once you've got a position there's nothing the carrier can do in those 30 seconds before impact so you won't hit it, so there is not even a necessity for terminal course correction.

link
nradov 1682 days ago
Total science fiction. You're not going to see enough from orbit with the type of optical sensor that can fit on the re-entry vehicle.
link
sudosysgen 1682 days ago
Actually yes, you can. You can run the numbers if you wish. To see a target that is 300m large from 300km away you need an angular resolution of 1 miliradian which requires a 0.5mm aperture in the visible spectrum. To be able to image it well and tell it apart from something else consistently and obviously you'd have more than you need with a hundred times the resolution which would require a 50mm aperture.

In other words, absolutely and completely feasible. You could hold a camera with an 1200mm f/20 Cassegrain lens that could take a picture of a carrier from 300km with a fair amount of detail in your hands, let alone fit it into a 3 ton reentry vehicle.

link
nradov 1682 days ago
That's idealized theoretical optics. Not something that works through a real world atmosphere. A total joke.
link
sudosysgen 1682 days ago
I assure you this is actually how it works. Im not talking theoretical either, you can go out and take a picture of the International Space Station throught the atmosphere using equipment that can be held with a single hand.

You'll be able to see the solar panels and various other little details.

Now that's a 100 meter target 400-600 km away. An aircraft carrier is even easier.

See for example https://www.amateurastrophotography.com/how-to-photography-t...

It turns out the atmosphere is not that big of an issue for such a large target. The ISS is much worse because the atmosphere is at the beginning of the light path so diffraction is minimised, in reality a picture of an aircraft carrier from orbit with the same lens would be much sharper.

Also, in real life these kinds of optics are really close to the theoretical limit, just fyi. The theoretical limits become more difficult to reach for larger apertures and higher focal ranges, but that's a few orders of magnitude away from what's needed here.

link