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

Is stacking really needed for moon images?

  • Please log in to reply
20 replies to this topic

#1 descott12

descott12

    Viking 1

  • -----
  • topic starter
  • Posts: 673
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 15 March 2019 - 08:01 PM

Hello,

I have been experimenting with imaging the moon. I have been using SharpCap to capture a .ser, stacking with Autostackert, then wavelet sharpening in Registax. The results are really great but, on a whim, I decided to just capture a single frame and go directly to Registax for sharpening. The results were pretty similar between the two, at least to my untrained eye.

 

So, I am think that stacking a movie file may  not be needed. I realize that stacking is really important for jupiter, mars, saturn but maybe not for the moon, possibly because it is so bright and you can get a pretty great shot with a very short exposure.

 

What do you think? Am I way off?

Thanks



#2 overnight

overnight

    Vanguard

  • *****
  • Posts: 2203
  • Joined: 10 Jan 2017
  • Loc: San Diego, CA

Posted 15 March 2019 - 08:05 PM

Post the results you are comparing, since that would give a better idea of what you're referring to.

 

Stacking on the Moon will be a benefit due to a higher SNR and a larger amount of recovered detail.


Edited by overnight, 15 March 2019 - 08:12 PM.

  • bobzeq25 and kellyvictoria like this

#3 james7ca

james7ca

    Fly Me to the Moon

  • *****
  • Posts: 6791
  • Joined: 21 May 2011
  • Loc: San Diego, CA

Posted 15 March 2019 - 09:47 PM

I've gotten pretty good lunar images using a single frame from a mirrorless APS-C camera. However, for the best results you really do want to use stacking since that can help to overcome issues with your seeing conditions and seeing is by far the biggest problem when doing planetary (or lunar) imaging. There is a reason why these techniques are known as "lucky imaging" and you are more likely to get lucky ten percent of the time (say, the best 100 out of 1000 frames) than you are by taking one frame (at random) during that same period of time.


  • RedLionNJ and t_image like this

#4 t_image

t_image

    Mercury-Atlas

  • -----
  • Posts: 2850
  • Joined: 22 Jul 2015

Posted 15 March 2019 - 11:58 PM

Hello,

I have been experimenting with imaging the moon. I have been using SharpCap to capture a .ser, stacking with Autostackert, then wavelet sharpening in Registax. The results are really great but, on a whim, I decided to just capture a single frame and go directly to Registax for sharpening. The results were pretty similar between the two, at least to my untrained eye.

 

So, I am think that stacking a movie file may  not be needed. I realize that stacking is really important for jupiter, mars, saturn but maybe not for the moon, possibly because it is so bright and you can get a pretty great shot with a very short exposure.

 

What do you think? Am I way off?

Thanks

So the principle is "using multiple images to extracting more information than is available in any single image,
then processing to reconstruct towards a more ideal sense of your frame."

 

So I always thought that the DSO stacking to pull data out of the noise floor by stacking to increase integrated time and thus the SNR,

was radically different than

the solar system method ('lucky imaging') that increases the possible spatial resolution of an image that otherwise has a hard ceiling,

not by noise (as in faint DSOs), but by atmospheric distortion...

DSO imagery it seems averages/blurs out seeing issues, but I figure if there were sensitive enough sensors with bright DSOs, that "lucky imaging" would reveal benefits as well.

So the distinction with the planets v. Lunar doesn't follow as there is not a noise floor issue, but a distortion one .....

Although to be precise there are interesting elaborate processes that can be done with DSOs in PixInsight, etc...

 

With the Moon depending on your atmospheric "seeing' conditions,

and your optics spatial resolution (FL to sensor pixel pitch)......

Noticing the difference with ideal processing v. single shot may depend on the resolution details you are looking at....

And maybe for your requirements it might not be worth the effort of how you process.

 

An objective test of the value would be if a terrestrial resolution target was used to demonstrate the difference between momentary seeing and what good processing using multiple frames can do, since one isn't at the mercy of what crater they are comparing, etc....

 

You might be interested in this, where "lucky imaging" is being done with realtime video, realtime.......

with great examples.

And daytime terrestrial is 'brighter' than the Moon, so the brightness issue is moot......

http://on-demand.gpu...ideo/S5352.html


  • dakinemaui likes this

#5 Tom Glenn

Tom Glenn

    Surveyor 1

  • -----
  • Posts: 1630
  • Joined: 07 Feb 2018
  • Loc: San Diego, CA

Posted 16 March 2019 - 12:38 AM

 Lunar doesn't follow as there is not a noise floor issue, but a distortion one .....

 

On the Moon there is actually both.  If you capture the entire Moon, or at least a large section of it at once, you span an enormous dynamic range, and if you expose properly so as not to clip the highlights, the regions near the edge of terminator are buried in shadow.  If you try to stretch the histogram on a single frame to bring out detail in this region, you get noise and quantization error from a lack of signal and bit depth.  If you stack 1000+ frames and do the same, you get much more detail because of increased SNR, and avoid much of the quantization error.  So the benefit really extends to both SNR improvement, as well as improved resolving power. 

 

To the OP....I agree with those above that say we need a bit more information about your workflow.  You can do quite well on the Moon in a single frame, although this is usually limited to situations with smallish image sales, such as presenting the entire Moon in one shot, no larger than 8 inches across in display size.  If you start increasing the image scale to achieve higher resolutions and display sizes, stacking has a huge benefit.  

 

Also, there is a difference between taking a movie of say, 1000 frames, ordering them by quality in AS!3, and then picking the best frame, versus just taking a single frame image and calling it done.  In the first situation, you can probably get a nice result from the single best frame of the series if the seeing was good.  In the second situation (just taking a single shot), it is a complete gamble.  But if you do take 1000 frames and order them by quality, and the top frames are pretty good, it makes sense to stack them for an even cleaner result.  I made an example of such a series that I did on a set of good data that I obtained in a post months ago (link below).  You can see that although the single best frame of the stack of 3000 frames (in this case a small example ROI near the crater Clavius) was of pretty good quality, the results got better upon stacking.  

 

https://www.cloudyni...more/?p=8992897


Edited by Tom Glenn, 16 March 2019 - 12:41 AM.

  • Jon Rista and t_image like this

#6 MalVeauX

MalVeauX

    Fly Me to the Moon

  • *****
  • Posts: 5219
  • Joined: 25 Feb 2016
  • Loc: Florida

Posted 16 March 2019 - 08:29 AM

Hello,

I have been experimenting with imaging the moon. I have been using SharpCap to capture a .ser, stacking with Autostackert, then wavelet sharpening in Registax. The results are really great but, on a whim, I decided to just capture a single frame and go directly to Registax for sharpening. The results were pretty similar between the two, at least to my untrained eye.

 

So, I am think that stacking a movie file may  not be needed. I realize that stacking is really important for jupiter, mars, saturn but maybe not for the moon, possibly because it is so bright and you can get a pretty great shot with a very short exposure.

 

What do you think? Am I way off?

Thanks

Hi Dave,

 

It depends on imaging scale.

 

With a small scale, doing a full disc or partial phase, a single frame can do the job. dSLR shooters do it all the time with one shot and they're not even close to having ideal sampling. But, at such a small scale, the seeing is not nearly as apparent, and with tiny apertures the bad seeing is barely resolved. Plus, the imaging train allows for a lot of light at these small scales, so you're working with very comfortable exposure values on the camera that allow for minimal gain, etc, which would lead to better signal to noise ratio, etc.

 

But, at large scale, high resolution imaging, you cannot do it in one shot. You must stack frames. And since the moon is virtually unchanged you can capture as many frames as you want, and stack hundreds or even thousands, to get the cleanest high resolution capture of features in high resolution. Add to this using narrowband filters to help combat seeing (such as long wavelength, red or IR wavelengths), or to help squeak out more angular resolution (shorter wavelengths, such as green, and a much higher demand on good seeing), and it just further influences what is needed to record in large scale.

 

For consideration, here's an example from the 13th at two scales for you to compare, also imaged with a C8 and a wee ST80 achromat:

 

https://www.cloudyni...arch-13th-2019/

 

I stack lots of frames primarily to remove noise and then also to have a really high signal to noise ratio for processing purposes. But mainly, again, I do it for the noise. And the results are the silky smooth results you see above yet with all the detail recorded.

 

Very best,


Edited by MalVeauX, 16 March 2019 - 08:35 AM.


#7 descott12

descott12

    Viking 1

  • -----
  • topic starter
  • Posts: 673
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 16 March 2019 - 10:25 AM

Hi All,

Thanks for all the responses. Very interesting stuff. Stacking definitely makes sense to me now for all the reasons stated. However, it is pretty amazing what one can do with just a single frame and a little sharpening in RegStax.  Here is an example that got me thinking about all this in the first place. Clearly it will not be winning any awards but it is pretty cool nonetheless.  And this was on pretty bad night as the moon appeared to be boiling and  it was taken with a color camera.  

Evo 8 + ASI294MC. Gain 59, Exposure 0.48 ms, single frame captured in SC and sharpened in RS6

Attached Thumbnails

  • Moon_00008-small.png

Edited by descott12, 16 March 2019 - 10:28 AM.


#8 descott12

descott12

    Viking 1

  • -----
  • topic starter
  • Posts: 673
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 16 March 2019 - 10:26 AM

And the sharpened version. Tons of pixelation on the edge but I am still impressed!

Attached Thumbnails

  • Moon_00008-RS6-small.png

Edited by descott12, 16 March 2019 - 10:28 AM.

  • t_image likes this

#9 RedLionNJ

RedLionNJ

    Gemini

  • *****
  • Posts: 3373
  • Joined: 29 Dec 2009
  • Loc: Red Lion, NJ, USA

Posted 16 March 2019 - 04:46 PM

Thanks for posting a sample of the data.

 

That's an extremely low-resolution image. At this kind of resolution, you may well not notice any appreciable difference in the outcome of the two techniques.

 

As others have suggested, imaging at a medium or high resolution will instantly show a difference.



#10 Tom Glenn

Tom Glenn

    Surveyor 1

  • -----
  • Posts: 1630
  • Joined: 07 Feb 2018
  • Loc: San Diego, CA

Posted 16 March 2019 - 07:13 PM

Dave, the images you have posted are only 300 pixels across, but your camera has a 4144x2822 sensor.  Why are these examples so small?  Even with the forum restrictions, you could post up to 1600x1200 with this camera easily.  On your scope, the image scale would be something like 0.5 arcseconds/pixel (at prime focus), which actually is capable of showing quite a lot of detail, and is definitely getting into the image scale where stacking is helpful.  At only 300 pixels across, even if there was noise in the images, it would be impossible for us to tell.  


Edited by Tom Glenn, 16 March 2019 - 07:13 PM.


#11 Tom Glenn

Tom Glenn

    Surveyor 1

  • -----
  • Posts: 1630
  • Joined: 07 Feb 2018
  • Loc: San Diego, CA

Posted 16 March 2019 - 07:21 PM

The other comment I have is your exposure.  Is 0.48ms really the exposure you used (as in 1/2000s)?  There's no need for an exposure that short, unless you were capturing an ISS transit of the Moon.  You can increase SNR per frame on the Moon by increasing the exposure and lowering the gain.  Although I haven't used this camera, so I'm not familiar with the exposure settings or the gain scale.  If you set gain to 0, what type of exposure is required to fill the histogram?  



#12 descott12

descott12

    Viking 1

  • -----
  • topic starter
  • Posts: 673
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 17 March 2019 - 09:07 AM

Hi Tom,

Thanks for the comments. Yes the original capture was at 4144x2822 bin 1. For some reason my original image is 1869x1365 - I must have resized it. But it was still way too large (1.2 MB) to upload. And the file with same resolution after wavelet processing in RS6 was 7.3 MB.

 

The images I posted were just a small crop of an area that I thought demonstrated my point.

Regarding the exposure: Here is the SC settings file:

 

Binning=1
Capture Area=4144x2822
Colour Space=RGB24
High Speed Mode=Off
Turbo USB=80(Auto)
Flip=Vert
Frame Rate Limit=Maximum
Gain=59
Exposure=0.00048

 

I have never tried using a gain of 0. I will see how that works. Thanks for your help and interest.



#13 descott12

descott12

    Viking 1

  • -----
  • topic starter
  • Posts: 673
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 17 March 2019 - 10:14 AM

Small detail I forgot to mention: this image taken at f2 through my Hyperstar - the would explain why the exposure was so short.



#14 aeroman4907

aeroman4907

    Viking 1

  • -----
  • Posts: 867
  • Joined: 23 Nov 2017
  • Loc: Castle Rock, Colorado

Posted 17 March 2019 - 10:30 AM

Small detail I forgot to mention: this image taken at f2 through my Hyperstar - the would explain why the exposure was so short.

Hi Dave, your clarification of imaging at f/2 also explains some things.  If you wanted to critically sample the moon with your camera, you should be imaging closer to f/23.  You are way undersampled and I wouldn't use the hyperstar setup unless that is the only way you wanted to configure your scope.  Also at such a fast focal ratio, I would imaging achieving good focus would be extremely hard as well.


  • RedLionNJ likes this

#15 Tom Glenn

Tom Glenn

    Surveyor 1

  • -----
  • Posts: 1630
  • Joined: 07 Feb 2018
  • Loc: San Diego, CA

Posted 17 March 2019 - 10:44 AM

Yeah, so that changes everything!  I didn't realize we were talking about imaging at f/2!  The problem is that on a telescope, you change the focal ratio by altering the focal length, not the aperture (most typically).  So while your 8 inch SCT has a native focal length of around 2000mm, the new focal length is around 400mm.  Even here, stacking will have benefit, but your resulting image will be more on par with the image scale achieved from a DSLR and a medium telephoto lens.  Also, if you want to post a larger image, you can always create a 1600x1200 crop of the original, and then use compression to bring it under 500kb.  Although now that we know about your imaging at f/2, this explains the tiny size.....and the short exposure.  I would recommend not imaging the Moon with the Hyperstar system, but I can understand it's a pain to change the configuration of the scope....even though the Hyperstar system is advertised as being easy to change, I can imagine that in reality it's a pain.  



#16 descott12

descott12

    Viking 1

  • -----
  • topic starter
  • Posts: 673
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 17 March 2019 - 01:15 PM

Yes, I was looking at DSO's that night and the moon looked nice, so I swung over to it and took few shots.  I now have a 178MM that I use at prime focus for the moon at sun. Speaking of which...the sun is out and there is blue sky for the first time in weeks...



#17 BKBrown

BKBrown

    Skylab

  • *****
  • Posts: 4372
  • Joined: 22 Aug 2009
  • Loc: Northern Virginia, USA

Posted 18 March 2019 - 09:15 PM

Unless you just want a snapshot plan to stack, it is the only way to get the best results short of being in lunar orbit...

 

EdgeHD 11, red filter, stacked:

 

Albategnius mosaic

Albategnius mosaic_v1.jpg

 

Aristoteles

Aristoteles D9.jpg

 

You will have much better data to use in post-processing if you stack...

 

Clear Skies,

Brian snoopy2.gif


  • aeroman4907 likes this

#18 dakinemaui

dakinemaui

    Vostok 1

  • *****
  • Posts: 169
  • Joined: 14 Mar 2019
  • Loc: Colorado Front Range Area

Posted 21 March 2019 - 09:32 AM

I now have a 178MM that I use at prime focus for the moon at sun.

With those sweet 2.4 um pixels, F/12 is about where you need to be to critically sample (or better) the passband of that camera. Using a red filter, you're looking at F/8.



#19 Phillysoc

Phillysoc

    Explorer 1

  • -----
  • Posts: 55
  • Joined: 12 Mar 2019
  • Loc: Maryland

Posted 25 March 2019 - 12:00 PM

Very good advice on this thread.  Thank you



#20 TareqPhoto

TareqPhoto

    Skylab

  • -----
  • Posts: 4193
  • Joined: 20 Feb 2017
  • Loc: Ajman - UAE

Posted 25 March 2019 - 06:58 PM

My "Bad" "Stacking" moon results

 

23-52-45-lapl3-ap109.jpg

 

1.jpg

 

03-03-16-lapl3-ap193.jpg

 

Only from the last year, and i am a beginner.


  • Wouter D'hoye and cmooney91 like this

#21 Phillysoc

Phillysoc

    Explorer 1

  • -----
  • Posts: 55
  • Joined: 12 Mar 2019
  • Loc: Maryland

Posted 25 March 2019 - 10:27 PM

Yes, I was looking at DSO's that night and the moon looked nice, so I swung over to it and took few shots.  I now have a 178MM that I use at prime focus for the moon at sun. Speaking of which...the sun is out and there is blue sky for the first time in weeks...

Rain and clouds are not our friends.




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






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics