Removing Satellite Trails Even With Few Frames: Don't Rush to Delete the Whole Frame

You finish integrating and find a satellite trail across the frame, and many people’s first reaction is to delete the single frame with the trail. But if you shot few frames to begin with, throwing away a whole precious frame for the sake of one line is a real shame. The photons each pixel collected are precious — don’t delete a single frame that’s mostly good pixels just because of a few satellite trails. This article is about how, when you have few frames, to remove satellite trails without deleting a single one.

Why Rejection Fails When You Have Few Frames

The root of the problem is this: when you have too few frames, defects like satellite trails, dust, and dirt spots are hard to eliminate by “picking out outliers” during integration alone. Outlier rejection is a statistical method, and when the sample (the frame count) is too small, the statistics are unreliable.

Here’s a concrete example: only 8 frames, shooting NGC 4038 & 4039, one of which has a satellite trail. In this situation, none of the sigma-clipping-related outlier rejection methods can effectively remove that trail. The traditional approach is to just delete that frame, but 8 frames is already few enough — deleting one hurts even more.

Two Roads: Adjust Parameters, or Force Rejection Manually

When removing satellite trails with few frames, there are two roads you can take.

Road one: adjust the rejection parameters. First try to dial in the rejection algorithm and parameters. For example, one case had 6 single frames, one of which had two bright satellite trails; after choosing an appropriate rejection algorithm and parameters (this example used percentile clipping), the trails were completely eliminated after integration. If this road works, it’s the least trouble.

Road two: force rejection manually. If the trails still won’t go away after adjusting parameters, you can use a “manual forced rejection” approach — draw a black line directly over the satellite trail to force those pixels to become outliers, so they’ll be rejected during integration. (Why draw a black line and not a white one? Think about the direction of outlier rejection — I’ll leave that for the reader to ponder.) I recorded a demonstration video showing how, with only 8 frames and sigma clipping doing nothing, you can use this manual approach to forcibly remove that trail without deleting a single frame (demo video: https://youtu.be/iQArxxS_K3Y).

Updating the Mindset: Old Rules Can Be Broken

Manually drawing a black line like this steps on a “taboo” found in plenty of teaching material. Many books or tutorials put it this way:

1. The original RAW files you shot must not be modified before integration.
2. The white point at the far right of the levels must not be touched, because it kills data.

These are actually fairly old processing conventions. The point is that the user needs to know what they’re doing — as long as you understand the principles, you can naturally draw a black line over a satellite trail to force it into outlier rejection, and you can also adjust the white point to brighten the cores of some faded-out stars. The rules exist to keep beginners from accidentally harming their data, but once you understand the mechanism behind them, you can reasonably break them in the right places.