Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Can't remove Background without losing all of M51

Astrophotography Beginner DSLR Refractor
  • Please log in to reply
20 replies to this topic

#1 ForestWise

ForestWise

    Lift Off

  • *****
  • topic starter
  • Posts: 9
  • Joined: 16 Nov 2020

Posted 16 April 2021 - 11:32 AM

Hi All, 

 

I was testing my imaging gear last night and I ended up with this image of M51 from my backyard in Pomona CA, (Bortle 9 zone)

 

Gear:

Mount: iEXOS-100 PMC-8

Camera: Canon M50

Telescope: Celestron Firstscope80eq (ST80 clone)

Guide Cam: ZWO ASI 120mini

Guide Scope Svbony sv165 (30mm/120mm F/4)

Guide Software: PHD2

Camera +Mount Control: NINA

 

I used

20x120s subs (40min total integration time) ISO800

6 Darks

30 Flats, taken with telescope against laptop monitor at max brightness on white screen

30 Bias

 

My issue right now, is that I can't seem to remove the background at all without it taking all of M51 away with it. Additionally, there doesn't seem to be any data in the Red and Blue channels when I change up the curves and edit by color.  In this case, would the best thing to do just be to add more integration time to the image and hope that the galaxy brightens up enough to not be removed with background exlusion?

 

All advice is appreciated, thanks!

 

Attached Thumbnails

  • Screenshot 2021-04-16 093425.jpg

Edited by ForestWise, 16 April 2021 - 11:35 AM.

  • limeyx likes this

#2 acommonsoul

acommonsoul

    Ranger 4

  • -----
  • Posts: 382
  • Joined: 08 Jan 2014
  • Loc: Hacienda Heights, Ca

Posted 16 April 2021 - 11:37 AM

Hi All, 

 

I was testing my imaging gear last night and I ended up with this image of M51 from my backyard in Pomona CA, (Bortle 9 zone)

 

Gear:

Mount: iEXOS-100 PMC-8

Camera: Canon M50

Telescope: Celestron Firstscope80eq (ST80 clone)

Guide Cam: ZWO ASI 120mini

Guide Scope Svbony sv165 (30mm/120mm F/4)

Guide Software: PHD2

Camera +Mount Control: NINA

 

I used

20x120s subs (40min total integration time) ISO800

6 Darks

30 Flats, taken with telescope against laptop monitor at max brightness on white screen

30 Bias

 

My issue right now, is that I can't seem to remove the background at all without it taking all of M51 away with it. Additionally, there doesn't seem to be any data in the Red and Blue channels when I change up the curves and edit by color.  In this case, would the best thing to do just be to add more integration time to the image and hope that the galaxy brightens up enough to not be removed with background exlusion?

 

All advice is appreciated, thanks!

Hi neighbor (Hacienda Heights here). What program are you using to extract the background?


Edited by acommonsoul, 16 April 2021 - 11:37 AM.


#3 N1ghtSc0p3

N1ghtSc0p3

    Viking 1

  • *****
  • Posts: 513
  • Joined: 19 Feb 2011
  • Loc: Texas

Posted 16 April 2021 - 11:38 AM

Forest, 

 

What application did you use to process your data? Can you post a master file somewhere (GoogleDrive)? It looks like a decent integration (no double stars or star trails or anything), but without knowing the steps you took, and in what application, it can be very hard to diagnose problems. 


  • Astrolamb likes this

#4 Astrolamb

Astrolamb

    Vostok 1

  • -----
  • Posts: 159
  • Joined: 26 Feb 2020
  • Loc: Arlington, Texas

Posted 16 April 2021 - 11:51 AM

I've gotta agree with N1ghtSc0p3, without seeing the unedited stack as a tiff or fits file this is hard to diagnose.

Also depending on what software you are using for stacking and post processing there are some settings that will give you this very wild Green color. 

I've had it happen on DSS.. 



#5 wrnchhead

wrnchhead

    Mercury-Atlas

  • *****
  • Posts: 2,834
  • Joined: 28 Aug 2017
  • Loc: NE Kansas

Posted 16 April 2021 - 12:11 PM

Another thing, it's all about signal to noise. The light pollution, is noise. At my bortle 6-7 home, I need about 6 hours total integration to get a decent result. 


  • rj144 likes this

#6 Sky King

Sky King

    Messenger

  • *****
  • Posts: 431
  • Joined: 16 Mar 2017
  • Loc: Arizona

Posted 16 April 2021 - 12:13 PM

Tried Affinity Photo, ProDigital AutoFlat Pro plug in, HSV, and Channel mixer. But I'm just learning myself. I'm sure someone can do better.

 

redo.jpg



#7 rj144

rj144

    Messenger

  • -----
  • Posts: 484
  • Joined: 31 Oct 2020

Posted 16 April 2021 - 12:31 PM

I think more data would also help.  



#8 ForestWise

ForestWise

    Lift Off

  • *****
  • topic starter
  • Posts: 9
  • Joined: 16 Nov 2020

Posted 16 April 2021 - 01:54 PM

I have been using GIMP to edit the photo. 

 

First I stretched gently several times, and that was all I got to before I got stuck. 

 

I stacked in DSS, and i've been using the Canon RebelTXi debayering routine.  Otherwise stock settings in DSS.

 

Here's the stacked tif: https://drive.google...iew?usp=sharing

 

since it's clear tonight and tomorrow night, I'll probably try to add a couple more hours of integration time. 



#9 unimatrix0

unimatrix0

    Mariner 2

  • -----
  • Posts: 209
  • Joined: 03 Jan 2021

Posted 16 April 2021 - 02:19 PM

 

Telescope: Celestron Firstscope80eq (ST80 clone)

 

 

I used

20x120s subs (40min total integration time) ISO800

6 Darks

30 Flats, taken with telescope against laptop monitor at max brightness on white screen

 

First of all, the flats didn't work, and it won't work like that.  You need to look at the histogram, once you do a test shot, with DSLR or mirrorless you want your histogram showing just around or beyond the middle line. The way it worked for me with DSLRs (or mirroless) is to switch it to aperture priority mode , take 1 image and look at the histogram and reduce or brighten the screen according to the histogram. 
Your image shows vignetting, which is due to the lack good calibration frames. Also 6 darks are not enough. You should have taken at least 20 if you ask me. 
And last, that scope is F11 which is considered slow, so it lacks on light gathering. There are F11 imaging scopes, mainly catadioptric high magnification, but those guys do like 4-5 min per subframe. 
Here is a stretched and somewhat color corrected snapshot of your stack.
 

Attached Thumbnails

  • snippi.JPG


#10 Astrolamb

Astrolamb

    Vostok 1

  • -----
  • Posts: 159
  • Joined: 26 Feb 2020
  • Loc: Arlington, Texas

Posted 16 April 2021 - 02:57 PM

I didn't go crazy or anything editing since this is mainly about the colors.. 

I used GIMP for all of the editing. 

 

After opening the file in GIMP

  1. I stretched the picture with the "Levels" tool in the "Colors" drop down 
  2. Opened the "Channel Mixer" tool by going into the "Colors" drop down and Hovering over "Components" near the bottom of the drop down. 
  3. Added about 0.2 from blue and 0.5 from green into the red channel and boosted the red in red channel to about 1.5
  4. Added about 0.3 from red and 0.3 from blue into the green channel and boosted the green in green channel 1.2
  5. Added about 0.2 from red and 0.6 from green into the blue channel and boosted the blue in blue channel to about 1.6
  6. Pressed okay and then duplicated that layer
  7. On the duplicated layer I ran a "Despeckle" from the "Filters" drop down within "Enhance" 
  8. I used the clone tool to take the area closest to the galaxy without any galaxy in it and painted that over the galaxy with the gradient left by the "Despeckle"
  9. I changed the duplicated layer to subtract and applied a Gaussian blur that was just high enough to get rid of the remaining gradient that was left by the "Despeckle"
  10. I then cropped the image down to get rid of the vignette  
  11. Made a mask that I could smooth out the background noise with.
  12. Applied a light sharpen using a mask 
  13. Tweaked the colors a little to be a little more accurate
  14. Boosted the saturation a little 

Hopefully that helps! The good news is that there is nothing wrong with the stack or settings other than bad flat frames. More integration time will help a lot to catch more details.

 

Autosave


#11 ForestWise

ForestWise

    Lift Off

  • *****
  • topic starter
  • Posts: 9
  • Joined: 16 Nov 2020

Posted 16 April 2021 - 03:23 PM

But I'll be sure to redo the flats while looking at the histogram. I already had it in AV mode, so that's kinda weird.

I'll probably add another 10-15 or so Darks tonight and hopefully another couple hours of integration time.

 

Is it possible that the vignetting is caused by the telescope focuser drawtube only being 1.25in? I didn't replace the focuser on this telescope.

 


And last, that scope is F11 which is considered slow, so it lacks on light gathering. There are F11 imaging scopes, mainly catadioptric high magnification, but those guys do like 4-5 min per subframe. 

 

This scope is a 80mm/400mm F/5, and I don't think they sell it as an EQ telescope anymore. I think they've since repackaged the OTA as the Celestron Travel scope 80, based on the specs. But it's light gathering capability is semi hampered at only 80mm anyways.

 

I can live with the CA for now, and I'll probably grab an APO or ED doublet when I have a chance to upgrade my mount. 



#12 N1ghtSc0p3

N1ghtSc0p3

    Viking 1

  • *****
  • Posts: 513
  • Joined: 19 Feb 2011
  • Loc: Texas

Posted 17 April 2021 - 09:24 PM

Hey Forest,

 

Thanks for posting the .tiff file. I put it into PixInsight and tried some basic processing steps:

 

Cropped the original file

Rotated 90 Degrees Clockwise

CanonBandingReduction Script

Rotated back 90 degrees counterclockwise

DynamicBackgroundExtraction

AutomaticBackgroundExtraction

Color Calibration

MultiScaleLinear Transform (noise reduction)

SCNR (to remove green tint)

TGVDenoise (more noise reduction)

And then some adjustments of the histogram to go from linear to non-linear. I'm not sure how this is handled in GIMP/PS, but in PI there's a function to "auto-stretch" it on the monitor only for all the linear processing steps (those listed above this sentence), and then you can transfer those settings to "Histogram Transformation".

 

Here's what I came up with:

M51_DBE_ABE.jpg

 

I think more subs would help smooth out the background and improve your SNR. This actually isn't a bad image...it just needs more data, and possibly a UV/IR block filter (which would help constrain the bloated blue stars.) You actually got a couple of faint fuzzies in there in the top right corner.

 

PI is expensive, but it is also very powerful and has a veritable chest of tools available to help with a LOT of different issues.

 

Hope this helps!


Edited by N1ghtSc0p3, 17 April 2021 - 09:39 PM.


#13 BQ Octantis

BQ Octantis

    Aurora

  • *****
  • Posts: 4,555
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 17 April 2021 - 10:34 PM

Beautiful!

 

Here's ten minutes in Photoshop:

 

(Click for full size @ 100% scale.)

gallery_273658_12412_141809.jpg

 

Yes, you need more data. I'd estimate a total of about ~10-15× what you captured (so 400-600 minutes total) for smooth, low-noise results at full scale.

 

If you get access to Photoshop, I'll share the workflow…

 

BQ


Edited by BQ Octantis, 17 April 2021 - 11:37 PM.


#14 GlendaleGuy

GlendaleGuy

    Vostok 1

  • -----
  • Posts: 150
  • Joined: 26 Sep 2017
  • Loc: Phoenix, Arizona

Posted 18 April 2021 - 02:19 PM

Hi All, 

 

I was testing my imaging gear last night and I ended up with this image of M51 from my backyard in Pomona CA, (Bortle 9 zone)

 

Gear:

Mount: iEXOS-100 PMC-8

Camera: Canon M50

Telescope: Celestron Firstscope80eq (ST80 clone)

Guide Cam: ZWO ASI 120mini

Guide Scope Svbony sv165 (30mm/120mm F/4)

Guide Software: PHD2

Camera +Mount Control: NINA


 

All advice is appreciated, thanks!

    This looks very similar to an experience I had when using a light pollution filter. But you don't state that is the case here. In my case I captured some more data without the filter and combined them all together in P.I. Currently I'm only using an IR cut filter.
 



#15 ForestWise

ForestWise

    Lift Off

  • *****
  • topic starter
  • Posts: 9
  • Joined: 16 Nov 2020

Posted 18 April 2021 - 03:22 PM

Hi Guys, thanks for all the help and sharing your workflows, I'm still working on adding more integration data to the image, i've got 82 lights (2hr 44min) and it's looking a lot better already, I'm planning on trying to get up to ~8 hours before calling it quits on exposure time. My neighbor leaving his backyard floodlight on has been cutting into my photography time though, so I only really get about an hour or two of time per night to image before teardown.

 

Do you guys think it's ok to leave this equipment out overnight to take photos, so I can get more subs per night and then do my teardown in the morning?

I'm still a beginner in terms of the photo editing side, and still learning the bits and pieces of photo editing beyond just taking the photo. Thanks for everyone that has shared their workflow.



#16 the Elf

the Elf

    Soyuz

  • *****
  • Posts: 3,516
  • Joined: 06 Sep 2017
  • Loc: Germany

Posted 18 April 2021 - 04:49 PM

You have quite some light pollution I guess. When you collect more data the image will become better. Here is what I can do with the data from your link:

 

m51.jpg

 

It is just not enough over all integration time.



#17 Sky King

Sky King

    Messenger

  • *****
  • Posts: 431
  • Joined: 16 Mar 2017
  • Loc: Arizona

Posted 18 April 2021 - 10:31 PM

A little more tweaking in Affinity Photo... (Affinity Photo software is still an amazing $25 to own.)  I'm not even doing it justice. This program is deep.

 

 redo3.jpg


Edited by Sky King, 19 April 2021 - 12:28 PM.


#18 BQ Octantis

BQ Octantis

    Aurora

  • *****
  • Posts: 4,555
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 19 April 2021 - 05:22 AM

I don't use Gimp much because Photoshop is a lot faster at most things—especially generating histogram previews when I'm testing out Levels settings. But it turns out, I can expand the Gimp histogram window way wider than Photoshop's, making its accuracy a fair bit better. Or at least the tradeoff is a better histogram preview for a slower update. So I tried my hand at a Gimp-only workflow. There's plenty of data even for Gimp—here's a 150% scale image with a Gimp-only workflow:

 

(Click for full size.)

gallery_273658_12412_214995.jpg

 

I'm nowhere near as proficient with Gimp as I am with Photoshop—I can anticipate the latter's behavior and know lots of simple shortcuts that I can't reproduce in Gimp. And I have zero equivalent to AstronomyTools 1.6 and Annie's Astro Actions in Gimp. So this is as far as I can get in 20 minutes.

 

This was all experimental, so I wrote nothing down. Let me know if you want me to do it again so I can document it…

 

BQ


Edited by BQ Octantis, 19 April 2021 - 05:22 AM.


#19 Schneebäuelli

Schneebäuelli

    Sputnik

  • -----
  • Posts: 29
  • Joined: 07 Jan 2020

Posted 19 April 2021 - 06:59 AM

Tried to do it with Pixinsight:

 

-RGB median align.

-Crop

-Multiple Dynamic Background Extractions

-CW 90° Rotation

-CanonBandingScript

-EZ-Denoise

-HistTranfrom

-Some CurvesTransfromations to reduce backgroundnoise visibillity.

 

 

Attached Thumbnails

  • Image05.jpg
  • Image05_2.jpg


#20 ForestWise

ForestWise

    Lift Off

  • *****
  • topic starter
  • Posts: 9
  • Joined: 16 Nov 2020

Posted 19 April 2021 - 06:49 PM

I'd like to thank everyone again for all your input, it's greatly appreciated, this is the image I'm at now, updated with way more data:

 

Now we have:

 

161 x 2min lights (322 min, 5 hours and 22 mins of lights!)

45 darks

30 bias

30 flats (redone)

 

I'm still thinking of adding more exposure time to this and changing up my editing a bit so I don't lose all my stars when subtracting my background. but it's looking way better now.

 

I'll post the stacked .tif if you guys would like to take a look at it. 

 

Attached Thumbnails

  • Screenshot M51.jpg

  • limeyx likes this

#21 BQ Octantis

BQ Octantis

    Aurora

  • *****
  • Posts: 4,555
  • Joined: 29 Apr 2017
  • Loc: Red Centre, Oz

Posted 20 April 2021 - 01:06 AM

Post away!

 

Still using Gimp?

 

BQ




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: Astrophotography, Beginner, DSLR, Refractor



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics