LocalNoramlization with "foreign" reference

[waelchlispixinsight]

Hi all

I have a set of frames taken with a 20" scope from a light polluted location. The frames show quite good resolution but - of course - severe gradients. After a lot of thinking and not little trial-and-error I came up with the idea to use local normalization with a reference image of the same object (same filter, different exposure time) taken from a very dark location but using a scope of only 1/3th the focal length of the 20"er. After registering and linear fitting (to one of the target frames to be normalized) of that reference frame I tried local normalization on one of the target frames. The result was a washed-out (smeared, very softened) galaxy and some sharp foreground stars. Obviously most of the frame content had been treated as background (at least to my understanding). A little of trial with the parameters did not yet solve the problem.

Any Idea if what I am trying to do is at all possible, leading to a good result,...? What is my misunderstanding otherwise?

Please do not advise me to redo the frames with less light pollution - this is not possible. The Instrument and the location are well worth some work as can be showed exemplary by an M1 image shot from there.:

http://www.upsky.ch/wp-content/uploads/2018/01/M1_3000mm_R1x1_39x5Min_G1x1_37x5Min_B1x1_35Min_Ha1x1_19x5Min.jpg

Thanks for reading, thinking and letting me know your findings.
Hansjörg

[waelchlispixinsight]

Found the solution: Must have been using the "file" of the reference image for the local normalization but applied the linear fit to the loaded view only (no save after linear fit). So the local normalization has been carried out towards an unfitted reference. Works fine now.