GOES Belt Satellite Trails: Best rejection?

[Terry Danks]

I am processing an image with numerous GOES trails. I have 22 lums and have selected Linear Fit Clipping as Rejection 1 in the Image Integration tool. No idea about Rejection 2, 3 or Large Scale Rejection settings. Have tried Winsorized Sigma as well.

Trails are substantially muted . . . but still present.

Is there a better way to suppress these evil things?

[pfile]

GOES as in the weather satellites?

the large scale rejection might be able to take care of them. another thing that rickS has mentioned is to open each sub and use pixelmath to draw a white line over each satellite trail, thus making them super-duper outliers and more easily rejected. i think adam block also recently posted a different method; it may have been based on LS rejection. can't remember offhand.

rob

[tloebl]

Just a thought....have you tried median combine instead of average? I assume the trails are not at exactly the same pixel locations on each sub.

[msmythers]

Terry

I would have a go with the new large scale rejection. Here is the original posting from Vicent. This particular image I would think would be of interest to you. It's an animated gif showing before and after. The other examples are great also.

https://pixinsight.com/forum/index.php?topic=11067.msg69188#msg69188

Mike

[RickS]

Definitely try Large Scale Pixel Rejection as Mike suggested.  As a last resort you can manually clobber the trails with PixelMath but it's a lot of work if there are many trails.

Cheers,
Rick.

[Terry Danks]

Oh wow! I thought I had multiple trails but the gif in Vicent's post makes my mess of trails look like "no big deal! looks that image was exposed through the rings of Saturn!

I tried the median vs average setting. Some improvement.

I am playing with large scale rejection now. Definitely getting better but I have not completely eliminated the trails as per Vicent's example.

Many thanks to all those who replied!

[RickS]

I'd treat median combine as a last resort.  It has a significant SNR penalty compared to an average combine.