Color Gradient

Reggie Jones

Hi Adam.  I'm working on Telescope Live data of NGC 4826.  The pre processing has gone ok so far until I completed PCC on an RGB combined image.  There is what I'll call a serious color gradient with the image.  I've used DBE several different ways and I'm still not getting a satisfactory result.  I even went back, used DBE on the individual LRGB integrated sub exposures and did RGB combination afterwards but the problem still persists.

I've been thinking of how to use Pixel Math to correct this, similar to how you did the single blip color removal, but I'm having trouble with what the expression might look like.  Suggestions?

Here's the link to DropBox with the file in question if you don't still have it.

https://www.dropbox.com/sh/gghsv15vdekzypo/AADsTGAiWmq8_SQ9UPO1ZMw0a?dl=0

Thanks

Reggie

BlockHead

In my quick and dirty attempt at this... I used DBE with a tolerance of 10... a smoothing of .1 and 15 samples per row (you can do 20 perhaps). Move any samples on or near stars away.. .shouldn't be too hard. This does take care of a good deal of it. Then I used SCNR on the Green channel. This gave me a relatively flat reddish background that is easy to correct and you will likely not show the image at this level. 

For the SCNR... you can protect things brighter than the background if you feel it necessary.

The attached screenshot looks reasonable to me. At the end of the day, I would not be displaying this background level. See what you think.

-the Blockhead

Reggie Jones

Wow, thanks for this; it's a lot better than what I came up with.  I'm going to need to do yet another review of your DBE module as I still don't quite fully understand how to use the Tolerance and Smoothing factors; I'd used 7 for Tolerance and .6 for smoothing.  I also used far fewer samples.

BlockHead

Reggie,

Since your "arcs" are on spatial sizes that are a small fraction of the frame, it makes sense to use less smoothing (meaning, the samples will be honored with a more aggressive fitting). It also makes sense to use the highest possible tolerance- since you know you can easily adjust samples and avoid stars.

That is the best logic I employed in my decision making.

-the Blockhead

Reggie Jones

Ok - your comments for smoothing and tolerance are really helpful; I'm still having some trouble understanding those parameters and how to use them effectively.  I've dropped what I've come up with so far after using your suggestions in DropBox.

Do you still do office hours?  I'd like about an hour of your time to ask some questions on DBE to make sure I finally understand how to use it properly and I have some questions around MLT and MMT.

Reggie

BlockHead

Reggie,

Before we consider 1 on 1 (which is tough for me right now...I am preparing for AIC among many other things)- did you review my sections on DBE, MLT, and MMT? This way, we are both on the same page with respect to questions. 

-the Blockhead

Reggie Jones

I have, more than twice.  Let me go through them once again (I view your tutorials usually more than twice anyway) and if I still have any questions, I'll just use this forum to ask them.

Thanks; R

BlockHead

Yeah... start here and ask some pointed questions like "you said this.. but this is malarky... " LOL

Of course I will have to listen to myself again.. but make a small list of questions here and lets see if we can work through some (and maybe benefit others that might have a similar question).

-the Blockhead

Reggie Jones

So, reviewed the DBE tutorials and now I think I get it; no questions there right now.

For MLT, I have 2 questions.  You used MMT to explain how the Threshold selection works for this particular module and you mentioned that you didn't quite know how the same selection actually works for MLT but you'd asked about it and were waiting for an answer - were you able to get an answer to this question?

Second question - For the Target selection at the bottom of MLT and MMT you can apply your selections to an RGB/K, Chrominance or Luminance; I'm not sure what the difference is between Chrominance and RGB/K.  Could you provide an explanation?

Thanks.  Reggie

BlockHead

Hi Reggie,

1. No... LOL... this was a while ago. I would need to review the video- but getting technical answers like this is very hard. There is just about ONE person who can answer- and he rarely does.

2. This deals with color space stuff and not my forte (meaning I do not know). There is a difference between the lightness of an image and the luminance. It is different math. Lightness and Luminance are slightly different things (calculated differently ... one retains hue, one does not). That is note the "restore CIE Y" is highlighted for one.

-the Blockhead