There are a surprising amount of false positives when using object detection on maritime imagery. While a carrier is a spec, there are a ton of specs in the ocean that can look similar enough. Plus, weather has a huge hand to play. If it were always perfectly clear, then it’s an easier problem, but one cloud can really mess up the detection. Ultimately, ship detection is a difficult problem (not intractable but still hard).
False positives are fine, you assign 1, 10, 50, 100 analysts to review hits. You only need to find it once, then the search area becomes incredibly small for each subsequent satellite pass.
I’m not saying that it is easy, just that you don’t need to have a surface ship within 15 nm in order to see it.
‘Take a picture of the entire ocean and look for ships’ is simple, but executing that plan is not.
It requires hundreds of millions of dollars of reconnaissance satellites, and an entire branch of personnel to operate and digest the information.
This is why the US operates carrier battle groups instead of just sailing their carriers everywhere with a small escort. They can’t hide, but they can pack enough offensive and defensive power into a tiny area to make most attacks infeasible.
Anyways, there’s a reason submarines exist
True, and even they’re vulnerable when they surface (if they’re moving), the v-shaped wake is also very detectable from space where satellites can detect wave heights within 3cm. It’s not easy for humans to find, but with billions of dollars to spend on computers, these kinds of things are very much within the reach of sovereign nations.
Seems like an simple but tedious job. Something that a computer can do.
Object detection algorithms are incredibly fast and can learn to tell the difference between an aircraft carrier and an ocean.
There are a surprising amount of false positives when using object detection on maritime imagery. While a carrier is a spec, there are a ton of specs in the ocean that can look similar enough. Plus, weather has a huge hand to play. If it were always perfectly clear, then it’s an easier problem, but one cloud can really mess up the detection. Ultimately, ship detection is a difficult problem (not intractable but still hard).
False positives are fine, you assign 1, 10, 50, 100 analysts to review hits. You only need to find it once, then the search area becomes incredibly small for each subsequent satellite pass.
I’m not saying that it is easy, just that you don’t need to have a surface ship within 15 nm in order to see it.
It kind of sounds like you’re saying that. Anyways, there’s a reason submarines exist
It is simple, it is not easy.
‘Take a picture of the entire ocean and look for ships’ is simple, but executing that plan is not.
It requires hundreds of millions of dollars of reconnaissance satellites, and an entire branch of personnel to operate and digest the information.
This is why the US operates carrier battle groups instead of just sailing their carriers everywhere with a small escort. They can’t hide, but they can pack enough offensive and defensive power into a tiny area to make most attacks infeasible.
True, and even they’re vulnerable when they surface (if they’re moving), the v-shaped wake is also very detectable from space where satellites can detect wave heights within 3cm. It’s not easy for humans to find, but with billions of dollars to spend on computers, these kinds of things are very much within the reach of sovereign nations.