alt Hacker NewsAveraging isn't the only option. It's possible to use other image-processing techniques which look at outlier values. This is way outside my area of expertise, but I believe sigma clipping is one of the standard go-tos, see:
<https://www.gnu.org/software/gnuastro/manual/html_node/Sigma...>.
More generally, you're clipping outliers:
<https://www.gnu.org/software/gnuastro/manual/html_node/Clipp...>.
This applies not only to satellite tracks, but meteors, cosmic rays, and other artefacts introduced into image capture. The techniques should be generally applicable, though for those who are specifically exploring transitory phenomena, this introduces additional challenges.
That’s an option. However, these satellites provide a predictable path so you don’t needed to detect them from image data. Which means you can even prevent them from showing up on long analog film exposures.