[finnx]

> approximately 4 × 10−4 of all pixels would be lost over the course of a year. However, simply counting pixels affected by satellite streaks does not capture the entirety of the problem, for example resources that are required to identify satellite streaks and mask them out or the chance of missing a first detection of an object

It looks like the main problem is not the amount of data lost but amount of extra manual work this situation causes. I assume Starlink tracks and knows where their satellites are, so why don't they just provide data feed to trusted third parties who might be affected by their satellites? That way researchers could automatically classify these trails.

[Frost1x]

I'd be surprised if the positions weren't already public data. Surely the US government requires or pressures knowledge of the positions since its a US based company and the satellites could certainly interfere with things NASA and other agencies want and need to do.

[japanuspus]

My understanding is that all startlink orbital data is available via https://www.space-track.org/.

[mlindner]

> It looks like the main problem is not the amount of data lost but amount of extra manual work this situation causes. I assume Starlink tracks and knows where their satellites are, so why don't they just provide data feed to trusted third parties who might be affected by their satellites? That way researchers could automatically classify these trails.

The information on where Starlink is is already publicly available through the US government. Anyone who wants to know where the satellites are can view the live positions at any time.

[jcims]

There have been quite a few folks build websites based on this data to show realtime state of the constellation:

https://satellitemap.space/

https://starlink.sx/

[tephra]

http://www.celestrak.com/Norad/elements/table.php?tleFile=st...

[SiempreViernes]

Starlink satellites are supposed to perform movements on their own, mainly to avid other satellites. But this means you might not know where they are all the time, just for the most part.

[Buttons840]

I don't know about that. The satellite must know where it's at to avoid other satellites. And if the satellite knows where it's at, why can't it tell us?

[capableweb]

> I assume Starlink tracks and knows where their satellites are, so why don't they just provide data feed to trusted third parties who might be affected by their satellites?

I'm not sure if this is in any way official and/or the right way of doing it, this area is all outside of my normal competence. But, stumbled upon a python library (https://github.com/python-astrodynamics/spacetrack) that supposedly connects to space-track (space-track.org) and you should be able to get the position there. How the data comes into space-track I'm not sure.

But there are bunch of small services for seeing the live location, so I'm sure someone is tracking the location somewhere, like this one: https://findstarlink.com/

[simondotau]

It seems to me that if you know where the satellites will be, it’s not so much a problem of removing streaks but rather factoring it into automated scheduling so that you never have any streaks to remove.