Color gradients appearing exclusively in mosaic overlap regions

Hi everyone,

I'm trying to understand and solve a persistent color gradient issue in my mosaics created with AstroPixelProcessor. I have performed quite a few tests with different APP settings, but the problem remains. I would appreciate any suggestions or explanations.

Equipment and acquisition details

• Camera: ZWO ASI2600MC-P25 (OSC)

• Lens: Samyang 135mm f/2@f2/4

• Acquisition software: ASIAIR Plus

• Exposure per panel: 10 × 60 seconds (10 minutes total)

• Gain: 100

All panels were captured during the same night. My main acquisition strategy is to capture each panel as close as possible to its local meridian, when the target is at its highest altitude and atmospheric extinction is minimized. I photograph the panels column by column, starting with the column currently crossing the local meridian. I capture all three rows of that column and then move to the next column, which has meanwhile reached approximately the same position in the sky.

This also helps keep the acquisition conditions and light pollution gradients relatively consistent across panels

The actual problem

After integrating the mosaic, I get visible green/cyan-colored regions exactly where the individual panels overlap.

Interestingly:

• The central regions of individual panels tend to have a slightly reddish/purple background.

• The overlap regions have a slightly green/cyan appearance.

• The color differences are not particularly obvious when looking at individual panels.

• The issue becomes visible mainly after combining the panels into a mosaic.

• The problem is most apparent in areas with very weak astronomical signal, such as relatively empty background sky.

• In regions containing the Milky Way or brighter nebulae, the green/cyan transitions are much less noticeable or practically invisible.

I have to apply a very aggressive stretch and saturation boost to make the problem clearly visible. With normal image stretching, these differences are much less obvious.

Here is an example

this is just a part of bigger mosaic from multiple nights

What I have already tried

I have experimented with different combinations of:

• MBB enabled/disabled/0-50%

• LNC enabled/disabled/1-4 order...

• Different normalization modes.

• Background calibration enabled/disabled.

The size and intensity of the green/cyan regions change depending on the settings, but they never disappear completely.

I also tried removing light pollution gradients from each individual panel BEFORE creating the mosaic. Unfortunately, the same green/cyan overlap regions were still present.

One particularly interesting test: I disabled both MBB and LNC. The green/cyan regions were still there, and their positions matched the overlap regions almost perfectly.

This makes me wonder whether the issue originates in the input panels themselves rather than in the blending algorithms.

Possible connection with different integration depths

Another aspect I am considering is the difference in effective integration time across the mosaic.

Each individual panel has only 10 minutes of integration.

However, an overlap region can(?) contain data from two, three or even four different panels.

Therefore, depending on the overlap geometry, the effective integration time could be:

• Panel center: 10 minutes.

• Two-panel overlap: 20 minutes.

• Three-panel overlap: 30 minutes.

• Four-panel overlap: 40 minutes.

Could the significantly different SNR between panel centers and overlap regions contribute to the appearance of these color transitions?

I understand that higher integration time alone should not change the expected RGB color ratios. However, I wonder whether small residual color differences between panels become more apparent when multiple panels are combined and random noise is reduced.

data:
https://drive.google.com/drive/folders/1DM-TKzqwNNFvSqLeQL7HPAHouXgDLJAB

Unfortunately, I am currently unable to log in to https://upload.astropixelprocessor.com/  

Thank you!