NGC 6992: Any PI Challangers for Heavy LP data?

[fatboy1271]

If anyone wants to try their hand at my NGC 6992 I'd love to see what you can come up with! I live in a Bortle 8-9.

I'm working my way through the Adam Block Studios (ABS) PI Fundamentals course. I think following the ABS BPP method should result in a good enough stacked image. That is how I PreProcessed the

XISF non-linear image I've uploaded:


CosmeticCorrection (CC)
BPP
• Bias, Darks, Flats
o Combination: Average
o Windsorized Sigma Clipping
o Sigma low = 3.5
o Sigma high = 2.5
• Lights
o Calibrate Only
 Apply CC
o Debayer
 RGGB
 VNG
• Global Options
o CFA images
o Optimize dark frames
o Up-bottom FITS
Hot Pixel Removal
• Defect Map
• PixelMath to Invert
• ImageContainer
StarAlignment
ImageIntegration

Data set is:
• Bias = 20
• Darks = 10
• Flats = 20
• Lights = 79 x 180sec at 1600ISO over two nights: 29 on 2019-09-17 and 50 on 2019-09-24
• Total Integration = 3.95

Equipment:
• AVX
• ES ED80
• Canon 70D
• ES 2" Field Flattener
• Hutech LPS-D1-48 (Light Pollution Filter)
• ASI120MC as Guide Cam
• Orion 50mm Guide Scope
• ZWO EAF Auto Focuser
• PoleMaster Polar Aligned

Software:
• PoleMaster
• CPWI
• PHD2
• NINA
• Stellarium

[Stuart]

Here ya go. Farted around with it while watching the last couple of innings of the Nats/Dodgers game

Automatic Background Extraction
Backgrond neutralizationb
Color calibration
Histogram transformation to make nonlinear
Curves for saturation through a lightness mask
SCNR
Another background neutralization (I didn't like the background color)
Morphologic transformation through a star mask to reduce star mass
ACDNR
Histogram transformation to adjust black point
Curves for contrast
Resample
.jpg

[fatboy1271]

Awesome, thanks Stuart!!! I appreciate you takin' the time to work on it. I mean, you were watching baseball so there's plenty of time to not be looking at the TV...

How was the data to work with? Is second BackgroundNeutralization normal or is it something to do with my captures?

Your Rock

fat

[Stuart]

Hey fat,

No, it's not normal to need a second background neutralization, and I'm not entirely sure it needed it. I just liked how it looked better.
In terms of the data, the color had a significant green cast and had a really weird color balance which was probably your light pollution. The SCNR helped. If I was less distracted I probably could have done more with it. You probably could benefit from longer subs if your system can handle it.

Stu

[fatboy1271]

Hey fat,

No, it's not normal to need a second background neutralization, and I'm not entirely sure it needed it. I just liked how it looked better.
In terms of the data, the color had a significant green cast and had a really weird color balance which was probably your light pollution. The SCNR helped. If I was less distracted I probably could have done more with it. You probably could benefit from longer subs if your system can handle it.

Stu

Hey Stu,

I believe the green cast is a combination of the LP itself and my LP filter. I should be able to do 300sec without issue; I've done 600 at 400ISO in the past. My fear has been how much the LP will effect those longer subs... I appreciate the effort, especially when something as important as a Playoff game is on. If you look at my signature you'll be able to tell that that isn't sarcasm!

fat

[Stuart]

Hey Fat,
You might try taking the LP filter off and seeing what you can get. Do an experiment in longer subs and then just look at the resulting histogram. If it's mostly on the left, then you're fine.

And yes, I love October. Although my team got crushed in their playoff game (A's)
Stu

[ram]

Here's my take on it (JPEG attached):

http://www.ram.org/images/space/downloa ... .vram.xisf

I did a DBE, AH, HT, CT, MLT and MT (morphological transformation). I'm not a big fan of this last process but in this case I felt it was aesthetically better.

I think your solution would have to involve a NB filter. I early on imaged with one of these nebula filters - they're like wide narrowband and they seemd to produce interesting data (which conflicted with my SHO data so I gave it up but I think they may work well with a colour camera).

--Ram

[yobbo89]

ok, i did a small histogram stretch, psf star profile,star mask for deringing and made a lum mask for the deconvolution of 25 iterations ,
scnr green reduction,curves tweak with saturation adjustment,tgvdenoise and then a further touch up with atrous wavelet transformation on the fine scale for slight blurring on the noise and a crop at the end

[yobbo89]

i've found something that's normally the trouble with this target. that big old bright star usually blows out on this target,yours is pretty good on star saturation, what settings did you use and exposure time ?

edit: nvm i found your details in the first post

[Baskevo]

I should be able to do 300sec without issue; I've done 600 at 400ISO in the past.

fat

How are you doing 600?? I can do about 200 at 800ISO, but that's with the histogram all the way on the right. Any further I'd blow out the histogram... Does the LP filter help that much?

The image looks great though! Nice shot fat, and great editing everyone!

[STEVE333]

First, thanks for sharing your data. I followed my normal workflow for this image.

While still linear I did BackgroundNormalization then ColorCalibration then used MLT for light noise reduction with an inverted luminance mask.

Stretched with HT
TGVDenoise with inverted luminance mask
Created starless version with Starnet
Boosted starless image color then used LHE to sharpen all through a luminance mask
Inverted luminance mask and reduced background of starless image
Inserted stars back into starless image using PixelMath
Reduced stars using MorphologicalTransformation through a star mask
SCNR to remove excess green (star mask removed)
S-curve to slightly reduce background while enhancing contrast
Boosted color saturation with Curves through a new luminance mask



Steve

[Stuart]

Those are some nice processes. Better than mine!

[fatboy1271]

Hey All!

Sorry for the late response; with a 7 month old and everything else in life I honestly forgot I posted this With that out of the way... WOW!!! You guys ROCK! @Stuart @ram @yobbo89 @STEVE333

I'll be looking your different processing lists!

Hey James, my 600sec exposures were done at 400ISO. As "proud" as I am to have gotten a good set of data at 600sec I am no longer taking subs that long due to the low ISO setting. When we go to a dark site I can try again

[Benjamin]

Stretched with HT
TGVDenoise with inverted luminance mask
Created starless version with Starnet
Boosted starless image color then used LHE to sharpen all through a luminance mask
Inverted luminance mask and reduced background of starless image
Inserted stars back into starless image using PixelMath

Hey Steve, I was just curious how you added the stars back in with Pixel Math? Star mask? I’ve just got Starnet++ up and running and curious how people use this tool.

[STEVE333]

Stretched with HT
TGVDenoise with inverted luminance mask
Created starless version with Starnet
Boosted starless image color then used LHE to sharpen all through a luminance mask
Inverted luminance mask and reduced background of starless image
Inserted stars back into starless image using PixelMath

Hey Steve, I was just curious how you added the stars back in with Pixel Math? Star mask? I’ve just got Starnet++ up and running and curious how people use this tool.

Hi Benjamin - I've just met a lovely lady and am starting a new chapter in my life. I mention this as an "excuse" for my late response.

After processing the starless image I use the following expression in PixelMath:

max(starless, original)

where starless is the name of the starless image and original is the name of the original image which was used with starnet++ to create the starless image. By the way, when processing the starless image I avoid reducing the background level. I wait until after I add the stars back in before reducing the background level.

Hope this helps,

Steve

[Benjamin]

Steve that’s lovely to hear on ALL fronts! Always appreciate those little details that can really make a difference :-)

[Benjamin]

So perhaps a process for adding RGB stars could be something like:
Apply Starnet++ to the RGB
Subtract the starless image form the original to give just the RGB stars
max(NBstarless, RGBstarsonly)
...although this depends on how ‘pure’ the RGB stars end up being?

[Benjamin]

Another option:
Apply starnet to the mono NB LUM
Subtract the Original NB LUM from the starless NB LUM, to create an NB star mask
Apply that NB star mask to the NB colour image and blend in the RGB stars (perhaps extracting L/a/b and blending the a and b channels to better keep the NB star profiles)

[STEVE333]

Another option:
Apply starnet to the mono NB LUM
Subtract the Original NB LUM from the starless NB LUM, to create an NB star mask
Apply that NB star mask to the NB colour image and blend in the RGB stars (perhaps extracting L/a/b and blending the a and b channels to better keep the NB star profiles)

So perhaps a process for adding RGB stars could be something like:
Apply Starnet++ to the RGB
Subtract the starless image form the original to give just the RGB stars
max(NBstarless, RGBstarsonly)
...although this depends on how ‘pure’ the RGB stars end up being?

Hi Benjamin - I've tried several different approaches to add the stars back into the starless image. So far, I've had the best success with creating an RGB image, running that image through Starnett++ to create a starless RGB image, enhancing the starless image (noise reduction, sharpening, increasing color saturation, etc.) then using Pixelmath with the max() operator to insert the stars back in.

Maybe one of your approaches will work better. Please let me know if you find a better approach.

Steve