I have used the d2line function in PixelMath a couple of times to clobber satellite trails that the rejection algorithm wouldn't clean up. I changed the values along the path in the offending sub to make them all 1.0 and so far from the average value that rejection removed them easily.
Cheers,
Rick.