Sharpening high resolution CaK images

this is the main message area for anything solar :)
Post Reply
LTHB
Im an EXPERT!
Im an EXPERT!
Posts: 433
Joined: Thu Mar 03, 2016 7:31 pm
Been thanked: 581 times

Sharpening high resolution CaK images

Post by LTHB »

Today I found time to process some CaK videos from September 19th. Seeing was quite good, but the sky was a little hazy, transparency not really good.

Working on these images I was, again, struggling with the question how much sharpening is appropriate if the aim is to depict the structures of the chromosphere as accurate as possible (neglecting, at least for the moment, any asthetic qualities of the images).

I finally came up with two versions:
Image

Image

The first one looks quite soft and blurry, so I applied more sharpening to get the second one. This one looks a little more crisp when scaled down to 75 percent:

Image

From the pictures posted in this forum (and the replies and comments here: viewtopic.php?f=4&t=24905) it seems to me that most people prefer more sharpened, crisp images (see e.g. the marvelous images here: viewtopic.php?f=4&t=24887, viewtopic.php?f=4&t=24940, viewtopic.php?f=4&t=25055).

If I take the images of the Dutch Open Telescope (operated at the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofísica de Canarias, http://www.staff.science.uu.nl/~rutte10 ... _home.html) as a standard, however, it seems to me that the first, blurry version of my picture is the one that comes closer to depicting the solar structures shown on the DOT images. Scaled down to compare them with the images above these look like this:

Image

Image

Image

My point is that, if the DOT pictures are reliable, then there are REAL soft, blurry structures to be seen in professional CaK images, which may be captured at least in part with our amateurs‘ high res equipment. These structures may be lost if our images are sharpened with the aim of getting a crisp look.

Now clearly a lot of unsharpness in my first image above is caused by imperfect conditions (hazy sky, quite high gain and therefore a quite large stack etc.) but comparing it to the (scaled down) DOT image I think it still might begin to capture some of the real blurred structures: In particular I think there are similarities in the light grey stuctures, their shape, variations of brightness, the variations of their density etc.

So it seems to me that the first version of my image is clearly unsharp (for the usual reasons) and surely could be improved upon, but it is an unsharp image of a soft, blurry structure, so sharpening it further to get something like the second version would remove some real structure from the picture, so to speak.

As always, comments and critique welcome!

Frank


User avatar
p_zetner
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 1644
Joined: Sun Mar 18, 2012 4:59 pm
Location: Winnipeg, Manitoba, Canada
Has thanked: 166 times
Been thanked: 719 times

Re: Sharpening high resolution CaK images

Post by p_zetner »

Hi Frank.

My guess would be that the DOT images you have posted have not had much, if any, processing applied to them. I'd say that most digital images, especially when oversampled, contain underlying details that can be made visible by deconvolution and / or sharpening. If you look at two of the scaled DOT images you posted (especially the second, middle, one) you can see "fibril-like" structures in some of the plage regions. To me, they're obviously a little blurred and beg to be enhanced by, at least, a small degree of sharpening! That implies, of course, that all the structures in the images are a little blurry and could also use some enhancement. No matter how big the telescope or how fancy the (adaptive) optics might be, there will always be some blurring introduced by the point spread function of the instrument (diffraction limit at the very least) and it's a legitimate exercise to try and ameliorate this with some digital image processing tools.

In the end, I suppose it depends on what you are after in the imaging. There's already a departure from a true rendition of the solar surface because of stacking! If you are after a pleasing visual effect that doesn't stray too far from reality then I think some deconvolution / sharpening is called for. I guess it's a bit of a judgement call to determine when an image is either underprocessed or overprocessed.

My two cents worth.
Peter


User avatar
Valery
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 4059
Joined: Sun Nov 18, 2012 3:13 pm
Has thanked: 156 times
Been thanked: 893 times

Re: Sharpening high resolution CaK images

Post by Valery »

DOT sun images are lucky imaging results. I can say more: the lucky imaging technique was basically developed by the DOT team.
This is the reason why all the lucky imaging amateur softwares (Registaks, AVIStack and AutoStekkert) are from the Netherland, where the DOT team is based.

This technique lucky imaging has been developed as a cheaper alternative to an Adaptive Optics.

The cadences of their CaK records are up to 40sec, typically 20-30 sec.

If to not too much concentrate on the time resolution, the lucky imaging is an "Adaptive Optics post factum". If an object does not changes too much during the recording cadence and the seeing has short moments of steadiness, then the results of AO and LI techniques are very comparable. We can see this on numerous examples of images of planets, the moon and the sun.
One can easily compare Hubble images of Mars with best images taken with ground based telescopes (Peak du Midi and Chiliscope).
and compare images of the sun taken with DOT telescope (lucky imaging) and Swedish Tower Telescope (which is AO) and NTT telescope at Big Bear Solar Observatory (AO).
More so, we can compare our own best images with the DOT images - they are quite similar (with reduction to the resolution, defined by the diffraction).

If one will ever buy/make a 16" scope with 16" DERF, install a scope at Canary Islands (or St, Veran, Peak du Midi), use a good CaK filter, his CaK images will be of the same level as DOT images.


Valery


"Solar H alpha activity is the most dynamic and compelling thing you can see in a telescope, so spend accordingly." (c) Bob Yoesle.

Largest full size 185 - 356mm Dielectric Energy Rejection Filters (D-ERF) by ARIES Instruments.
User avatar
Valery
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 4059
Joined: Sun Nov 18, 2012 3:13 pm
Has thanked: 156 times
Been thanked: 893 times

Re: Sharpening high resolution CaK images

Post by Valery »

TheSkyBurner wrote: Mon Oct 01, 2018 10:49 pm
1. user uploaded images online are a false impression of what is "real" and 99% of all astronomy photos are not "real";
They are an artist ability based on software processed enhancement.


1. Do not be fooled by user data processing.
I'd say:

1. 100% of raw astroimages are not real. A correct image restoration software gives on the output images which are much closer to the reality than raw unprocessed images.
For example, images shown here by Apollo Lusky, are suffer from:

a. telescope transfer function (diffraction, defocusing, miscollimation, wave front distorsion, geometrical field distorsion etc etc)
b. atmospheric blurring.
c. photon noise
d. camera dark noise and read noise
e. camera non-linear transfer function
i. camera pixels shape

Enough for conclusion that these images are anything, but not real.

2. Correct image processing can minimize most of these impact factors
and give us images which are much closer to the reality than raw ones.


Valery


"Solar H alpha activity is the most dynamic and compelling thing you can see in a telescope, so spend accordingly." (c) Bob Yoesle.

Largest full size 185 - 356mm Dielectric Energy Rejection Filters (D-ERF) by ARIES Instruments.
User avatar
Montana
Librarian
Librarian
Posts: 34560
Joined: Mon Oct 17, 2011 5:25 pm
Location: Cheshire, UK
Has thanked: 17667 times
Been thanked: 8787 times

Re: Sharpening high resolution CaK images

Post by Montana »

:) I wouldn't go so far as to say the camera never lies :) I was watching a programme the other day and they were filming a very unusual Swallow on the nest webcam. They had all the researchers at Bristol University trying to work out what bird it was because it was so unusual in colouration. They were using a colour camera and the Swallow was black and white, no red at all and all the background looked normal. After 2 days of researching they decided to go and actually look at the bird. It was a regular red, black and white Swallow. So the camera was very weird indeed and not functioning in the unusual way, and eyeballs were telling the truth. However, it is rather difficult to go and have a look at the actual Sun ;)

my rule of thumb is to watch the live image stream as carefully as possible and try to match what my eyes see in detail in the final processed image. The eye seems to see past bad seeing and see the detail.

Alexandra


User avatar
MalVeauX
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 1858
Joined: Tue May 09, 2017 7:58 pm
Location: Florida
Has thanked: 1171 times
Been thanked: 1360 times

Re: Sharpening high resolution CaK images

Post by MalVeauX »

LTHB wrote: Mon Oct 01, 2018 10:10 pm Working on these images I was, again, struggling with the question how much sharpening is appropriate if the aim is to depict the structures of the chromosphere as accurate as possible (neglecting, at least for the moment, any asthetic qualities of the images).

Frank
Hi Frank,

I think processing and the idea of sharpness are definitely a personal thing, but I found lots of "blurry" structures in CaK here and there while imaging at wider FOV, and I figured they were filaments that were masking the material behind them, but it wasn't super obvious unless we were sub-angstrom to see them clearly. Just a guess, I could totally be wrong there. But, to answer your question about how much sharpening is appropriate, under good seeing, without processing, I was staring at very obvious fine line pencil drawing structures. I think its too easy to over-sharpen. There's a delicate balance. But, I tried to mimic my processing of good seeing data to be as sharp as what I saw on my monitor while capturing the data with lucky imaging, as there were extremely well defined cells, pencil like drawing look to it, which is how I knew I was even in focus to begin with, and of course, really good seeing.

These are with a 120mm aperture in CaK and are 1:1 aspect crops:
cak_01.jpg
cak_01.jpg (125.21 KiB) Viewed 3701 times
cak_02.jpg
cak_02.jpg (173.22 KiB) Viewed 3701 times
cak_03.jpg
cak_03.jpg (301.64 KiB) Viewed 3701 times
After an initial very soft deconvolution, I sharpened via low value high pass filters, just enough to define the cell edges, and progressively worked to higher value high pass filters, generally 2, 3, 4, 5, etc, and that's it, then a wee bit of unsharpening mask at a very low pixel value and percentage.

I'm on the fence whether this is over-sharpened or not. It's similar to what I saw during good seeing on the screen.

Very best,


User avatar
krakatoa1883
Almost There...
Almost There...
Posts: 1041
Joined: Sat Jun 11, 2016 9:41 am
Location: Mediolanum
Has thanked: 153 times
Been thanked: 496 times

Re: Sharpening high resolution CaK images

Post by krakatoa1883 »

Deciding what processing shall be applied to calcium line images is not an easy task because it is not always clear how the chromosphere (at these wavelength) should look. Processing is a very subjective exercise to the point that most CaK (or CaH) images that one can see on the web are all somewhat different from one another (sometimes VERY different even if the equipment used is similar), it is therefore natural that one can ask which of them represent the "true" aspect of the Sun. Speaking for myself, for months I tried to obtain some decent results by using of IMPPG until I realized it generates too many artifacts.

I think however that there are some "reference" points that can be very useful in selecting what filters apply during processing and to which extent and these are for me the images by (among others) Christian and Peter. Those by Christian are always far beyond my reach :lol: but they set some sort of standard to which I think most solar imagers look at. Peter is making an excellent work too, I think for example to images like this one, closer to my reach because obtained through a ordinary 115 mm triplet. At a smaller scale these fine images match very well DOT's ones in this same spectral region, so for me they can be certainly assumed as a reference result for a skilled imager.

More specifically I was able to significantly improve my results by using of UM, (no RL deconvolution) and by the stacking of very few frames as suggested by Christian, no more than 15 - 20 out of about 1500 - 2000.


Raf
My solar images and reports with articles on solar equipment
LTHB
Im an EXPERT!
Im an EXPERT!
Posts: 433
Joined: Thu Mar 03, 2016 7:31 pm
Been thanked: 581 times

Re: Sharpening high resolution CaK images

Post by LTHB »

Thanks for all the responses and suggestions! Unfortunately I won‘t be able to really appreciate and digest them before the weekend, but then I will!

Till then I just wanted to share some reading material I found looking for information on the soft, blurry structures on the CaK images from the DOT: It seems this phenomenon is called „reversed granulation“ and there are some detailed studies of it, some based on a close analysis of the DOT images, e.g.:

http://iopscience.iop.org/article/10.1086/307219/pdf

https://folk.uio.no/svenwe/public/files/aa4115.pdf

http://www.staff.science.uu.nl/~rutte10 ... ..333R.pdf

http://www.staff.science.uu.nl/~rutte10 ... 1-p687.pdf

Looks quite interesting, and even potentially relevant for CaK imaging, as some of these structures (the "internetwork") seem to change on an time scale of less than one minute.

Thanks again and regards,

Frank


User avatar
marktownley
Librarian
Librarian
Posts: 42270
Joined: Tue Oct 18, 2011 5:27 pm
Location: Brierley Hills, UK
Has thanked: 20424 times
Been thanked: 10243 times
Contact:

Re: Sharpening high resolution CaK images

Post by marktownley »

Interesting thread. I think for me next years new toy will be a 115/7 triplet for my higher res CaK setup.


Image
http://brierleyhillsolar.blogspot.co.uk/
Solar images, a collection of all the most up to date live solar data on the web, imaging & processing tutorials - please take a look!
User avatar
marktownley
Librarian
Librarian
Posts: 42270
Joined: Tue Oct 18, 2011 5:27 pm
Location: Brierley Hills, UK
Has thanked: 20424 times
Been thanked: 10243 times
Contact:

Re: Sharpening high resolution CaK images

Post by marktownley »

Tee hee hee! That would be fun! :)


Image
http://brierleyhillsolar.blogspot.co.uk/
Solar images, a collection of all the most up to date live solar data on the web, imaging & processing tutorials - please take a look!
User avatar
krakatoa1883
Almost There...
Almost There...
Posts: 1041
Joined: Sat Jun 11, 2016 9:41 am
Location: Mediolanum
Has thanked: 153 times
Been thanked: 496 times

Re: Sharpening high resolution CaK images

Post by krakatoa1883 »

LTHB wrote: Tue Oct 02, 2018 11:07 pmthere are some detailed studies of it, some based on a close analysis of the DOT images, e.g.:
thanks, Frank, very much welcomed literature, it will keep me busy the next week end :D


Raf
My solar images and reports with articles on solar equipment
christian viladrich
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 2150
Joined: Sun Jun 14, 2015 4:46 pm
Location: France
Has thanked: 1 time
Been thanked: 2713 times
Contact:

Re: Sharpening high resolution CaK images

Post by christian viladrich »

Hello Frank,
Thanks for opening this very interesting question. This is one I tried to answer in the book I am publishing at the end of the month.
http://www.astronomiesolaire.com/
Indeed, I have a paragraph on "reverse granulation", and an extensive discussion on artifacts and the limit between processing and over-processing. So here is a sneak preview ;-)

Let's start with reverse granulation :
Reverse granulation lies at the bottom of the chromosphere, about 300 - 500 km over the photosphere. On the opposite of granulation, the interior of the cells are dark while the intergranular region is bright. The cells are larger compared to "normal" granulation and can be easily seen up to the solar limb :
http://astrosurf.com/viladrich/astro/so ... OA-ASI.jpg
Seeing "reverse granulation" is a good test of resolution, correct processing, and the sign of an instrument performing well in near UV.
As shown in the images above, it takes an aperture of about 100-120 mm to detect reverse granulation. The intergranular region is best seen with 150 mm aperture and beyond.

Processing CaK images :
There is a natural tendency in all of us of over-processing images when we want to show features which are the limit of detection. Back to your three images, I prefer your first one, since the other two show only increasing artifact.

Artifacts or actual features ?
This leads us to the very interesting question of being able to make the difference between artifacts and actual solar features.
Let's take a step backward before getting into this. There are three main approaches in astronomical image processing :
- aesthetical approach : the objective there is to have an image with deep visual impact. Artifacts are not an issue.
- realistic approach : this is an in between situation. The objective is to show actual features and no artifacts, and still have a "nice" image.
- specific approach : this one pushes the processing very far in order to show very low contrast details and the expense of getting large artifacts. One example is processing Uranus to show equatorial bands on Uranus using at strong wavelets and strong denoise. Usually, people using this approach are very familiar with the actual details and know the difference with artifacts.

Al three approaches are possible, this is just a question of personal choice.
If we consider the second approach, as a whole :
- white light images are the ones that are the easiest to process since everyone can easyly have an experience of what the solar looks like in white light. Furthermore the feature structure of the photosphere is very simple (granulation, solar spots).
- Ha imaging can more difficult since overprocessing (=artifact) might give additional "visual impact",
- Ca K imaging is definitely the more difficult since it takes good instrumentation to get actual details and high resolution images are more rare.

Avoiding artefact :

The first thing to do to avoid artifacts it to get familiar with the actual features on the sun. This can be done by :
- observing visually with the telescope (not an easy task in CaK ...),
- looking at images taken with larger apertures in the amateur world, or much larger apertures in the professional world,
- understanding the actual features on the Sun.
This is exactly the way you said at the beginning of the post.

Another complementary approach is to get an experience of high resolution imaging (moon, planets). The reasons is simple : if a processing method gives artifacts on lunar / planetary imaging, it will also gives artifacts on solar imaging. On the other way, if it gives no artifact on lunar / planetary imaging, it will give no artifact on solar imaging.
Basically, I have the same processing flow / processing settings for lunar / planetary / solar imaging. The main exception is Uranus imaging because the contrast of the bands is very low. This one falls in the "specific approach" described above.

The last point is to keep it simple. For example, avoid denoise, it is not needed in solar imaging. If the raw data is good, you don't have to apply fancy processing.

Sorry for this very long message ...Hope this helps


Christian Viladrich
Co-author of "Planetary Astronomy"
http://planetary-astronomy.com/
Editor of "Solar Astronomy"
http://www.astronomiesolaire.com/
User avatar
pedro
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 12256
Joined: Sun May 01, 2016 8:26 pm
Location: Portugal
Has thanked: 14 times
Been thanked: 6576 times
Contact:

Re: Sharpening high resolution CaK images

Post by pedro »

The book is very interesting Christian, I made a pre-order today


User avatar
Valery
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 4059
Joined: Sun Nov 18, 2012 3:13 pm
Has thanked: 156 times
Been thanked: 893 times

Re: Sharpening high resolution CaK images

Post by Valery »

christian viladrich wrote: Sat Oct 06, 2018 6:26 pm
This is one I tried to answer in the book I am publishing at the end of the month.
http://www.astronomiesolaire.com/
Indeed, I have a paragraph on "reverse granulation", and an extensive discussion on artifacts and the limit between processing and over-processing. So here is a sneak preview ;-)


Reverse granulation lies at the bottom of the chromosphere, about 300 - 500 km over the photosphere. On the opposite of granulation, the interior of the cells are dark while the intergranular region is bright. The cells are larger compared to "normal" granulation and can be easily seen up to the solar limb :
http://astrosurf.com/viladrich/astro/so ... OA-ASI.jpg
Christian, thanks for your input here.

BTW. When the english version of the book will be available?

BTW2 "can be easily seen up to the solar limb" this your definition "easy" is not easy at all for single stacked Lunt CaK, DayStar CaK, DayStar CaH and PST mod CaK. Only double stacked Lunt CaK and PST CaK allow comparable to your Barr filter performance at the sun limb.


Valery


"Solar H alpha activity is the most dynamic and compelling thing you can see in a telescope, so spend accordingly." (c) Bob Yoesle.

Largest full size 185 - 356mm Dielectric Energy Rejection Filters (D-ERF) by ARIES Instruments.
christian viladrich
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 2150
Joined: Sun Jun 14, 2015 4:46 pm
Location: France
Has thanked: 1 time
Been thanked: 2713 times
Contact:

Re: Sharpening high resolution CaK images

Post by christian viladrich »

Hello Valery,
I hope we have an English version of the book in one year or so.
I do agree, my use of "easy" was a bit of an overstatement ;-) My point was to say that, compared to visible light and "normal" granulation, we can see "inverse granulation" closer to the limb in Ca K.

Apollo, the sun is getting lower and lower, this is probably the end of the solar season for me.


Christian Viladrich
Co-author of "Planetary Astronomy"
http://planetary-astronomy.com/
Editor of "Solar Astronomy"
http://www.astronomiesolaire.com/
User avatar
Valery
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 4059
Joined: Sun Nov 18, 2012 3:13 pm
Has thanked: 156 times
Been thanked: 893 times

Re: Sharpening high resolution CaK images

Post by Valery »

christian viladrich wrote: Sun Oct 07, 2018 6:22 pm
this is probably the end of the solar season for me.
Same for me. However I hope for some warm higher pressure area will "visit" us this season.

But... as you said earlier, you have a lot of material (taken last summer) to process. Lets do the processing of these materials and show to our collegues here. Otherwise, what the use of these stocked materials? What do you think?


Valery


"Solar H alpha activity is the most dynamic and compelling thing you can see in a telescope, so spend accordingly." (c) Bob Yoesle.

Largest full size 185 - 356mm Dielectric Energy Rejection Filters (D-ERF) by ARIES Instruments.
User avatar
krakatoa1883
Almost There...
Almost There...
Posts: 1041
Joined: Sat Jun 11, 2016 9:41 am
Location: Mediolanum
Has thanked: 153 times
Been thanked: 496 times

Re: Sharpening high resolution CaK images

Post by krakatoa1883 »

christian viladrich wrote: Sat Oct 06, 2018 6:26 pmThanks for opening this very interesting question. This is one I tried to answer in the book I am publishing at the end of the month.
http://www.astronomiesolaire.com/
good to know, Christian, I pre-ordered a copy.

The grainy structure of the chromosphere in calcium light toward the limb can be imaged even with small apertures although not with the resolution of this picture, of course, however it can certainly be detected.

Agree with Valery that your Barr filter is truly exceptional.


Raf
My solar images and reports with articles on solar equipment
User avatar
p_zetner
Way More Fun to Share It!!
Way More Fun to Share It!!
Posts: 1644
Joined: Sun Mar 18, 2012 4:59 pm
Location: Winnipeg, Manitoba, Canada
Has thanked: 166 times
Been thanked: 719 times

Re: Sharpening high resolution CaK images

Post by p_zetner »

I’m really enjoying this thread!

Christian, I’ve placed my pre-order for your book.

Cheers.
Peter


Post Reply