Long exposures can still have a duration censored. This is digital photography, not chemical. If you expose for 1 million seconds, you can edit out 60 seconds where a satellite passes. That data is forever gone, but you still get the rest of the data from that pixel time.
Given that the earth rotates every 1.5 thousand or so minutes something tells me 1 million may not be representative...
Sure, I was being extreme, but multi day or week surveys exits. In fact, they already have to do a lot of filtering to remove transient artifacts
You're assuming there's only one satellite to worry about per exposure.
Maybe two telescopes and a bit of math to use the difference in distance to the satellites vs the asteroids?