Integrating Lum frames of a dim target, I have found some curious horizontal tracks or shadows. They appear to be correlated with the brightness of the star from which they emanate.
I've attached a screenshot.
The tracks aren't a readout artefact - they are far too fat for that and don't align with the way this chip reads out its data - that is at a 90 degree rotation from this.
I can't see anything on the individual frames so I concluded its coming from the integration process. That said, I find it odd that I have this in Lum but not RGB - perhaps because I take Lum unbinned and RGB binned x2. The camera is a STX16803.
I have played with many settings in the Image Integration tool and can't eliminate this. The only setting that improves it is to set the Combination setting from Average or Median to Minimum of Maximum - but then I dump plenty of data too.
Have you tried doing a really big stretch on a couple of the raw subs to see if you can see any traces of the dark tracks? Perhaps try inverting the stretched image as well. If there are no traces there, I'd try the same on some calibrated subs to see if the calibration process is causing the problem. Next thing I would try is a simple average integration without any rejection, maybe even without normalization. Just play around and look for where they are coming from. Having found the source you might be able to do something about it.
My guess is that the tracks are in the raw data but very faint.