The data for this image is about about 4 years old but the processing continues to evolve.
It is a 30 hr luminance exposure with a BRC-250 scope and ST-X10ME.
The surface background of the galaxy is much fainter than the natural skyglow at the darkest sites and is only "visible" as the surface background is added to the skyglow.
The first attachment is a conventionally processed image of the object.
The second attachment is skyglow subtracted which I developed a few years ago.
The third attachment is a pixel mapping procedure I developed in PI and is applied to the second attachment.
The pixel mapping curve for the procedure is shown in the final attachment.
Nicely done, Steven. What's the surface brightness of this beast? Just wondering how it compares to faint stuff like the 4th jet of NGC 1097.
Cheers,
Rick.
Thanks Rick
From the literature the surface brightness of the of the Carina Dwarf is the range of 25.7-33.2 mag/sq arc sec.
The problem is being able to accurately define the size of the Dwarf by being able to delineate it from our own galaxy.
Quote:
Originally Posted by Octane
Is that faint cirrus surrounding the dwarf now? :O
Unreal.
H
Thanks H.
It could be the tidal stream as the dwarf is being pulled apart by our galaxy.
The outer edges visible in the third image apparently extend into the tidal radius of the galaxy and has a surface brightness of around 33.0 mag/sq arc second.
The outer edges visible in the third image apparently extend into the tidal radius of the galaxy and has a surface brightness of around 33.0 mag/sq arc second.
Steven
G'day there, SJ,
if confirmed, that is a very impressive diffuse-object detection limit, even when compared to the detection limits of long exposures with big telescopes that are published in the professional literature of astronomy.
It would seem that you are also imaging fainter than most of the "ultra-deep" exposures that are increasingly finding their way into the IIS astro-imaging forum.
In the literature during the first decade of the 21st C. , I recall that some of the "pros" found it challenging, in their imaging of diffuse objects, to get an adequate signal-to-noise ratio even at a surface brightness of 28 V magnitudes per square arcsecond, so I am still in need of further convincing regarding your extremely faint surface brightness limit.
But if the various structures detected in your image of Carina are confirmed, it is consistent with the facts that you could well be doing better than everybody else at bringing out ultra-low-surface-brightness details in galaxies.
As yet, I have no absolute proof of this assertion of the superiority of your imaging and image-processing techniques, but I note that professional astronomers still have to struggle to get really clear images of the inter-galaxy (intracluster) Diffuse Optical Light within clusters of galaxies, which is known to be at 26.5 to 32 V magnitudes per square arcsecond.
Certainly, if you can image the very outermost regions of galaxies to these ultra-faint limits, and you can display the images at high contrast, there are going to be a lot of interesting and unusual detections in your images!
cheers,
Robert
P.S.
I am trying to think of some dwarf galaxy specialists who would be interested in your image and in your imaging techniques (I have them somewhere in my files). Also, some of the members of the 'Intracluster Light community' might be interested in your imaging techniques.
______________________________
Just "for fun and profit" :
Image of M86/M84/M87 region to 28 V magnitudes per square arcsecond.....
______________________
Last edited by madbadgalaxyman; 13-10-2014 at 09:09 AM.
Perhaps, if you really can image as faint as 33 magn. per square arcsecond, you could be doing surveys in order to discover ultra-faint dwarf galaxies which are so dominated by dark matter that their luminous component is composed of the merest smattering of faint stars: http://www.physics.mcmaster.ca/Fac_H...sityDwarfs.pdf
These extremely low-luminosity dwarf galaxies are the objects that are closest in their properties to the tiny "dark matter halos" that have been predicted in cosmological simulations.
The so-far discovered 'dark galaxies' do emit some extremely-extremely-low surface brightness light from a handful of constituent stars, but only a tiny fraction of the total mass of one of these galaxies is in the form of ordinary luminous matter.
_______________
Last edited by madbadgalaxyman; 13-10-2014 at 08:39 AM.
One way of dealing with the "authenticity" is to process the raw data of the ESO image and let ESO decide.....
This has been done in the past.
I'll send an E-mail to Lars Lindberg Christensen at ESO to request their raw data.
I've taken the Virgo Cluster images and applied my pixel mapping routine.
Attachment 1 is the original image.
Attachment 2 is 7 iterations of my pixel mapping routine to Attachment 1 revealing the diffuse light.
Attachment 3 is the original push processed image.
One way of dealing with the "authenticity" is to process the raw data of the ESO image and let ESO decide.....
If anyone is interested apart from Robert, the European Southern Observatory (ESO) through the astronomer Olivier Hainaut has graciously supplied me with raw images of the Carina Dwarf taken with the 2.2 metre MGP/ESO and 4 metre Victor M Blanco scopes in Sth America.
Olivier's activities have included ultra deep imaging on 6 and 10 metre scopes and extensive experience with image processing.
Since this activity is an exercise in science and not in the production of a pretty picture there are some strict guidelines.
(1) No sharpening
(2) No contrast enhancement.
(3) No colour saturation.
(4) No photo-shopping etc.
The raw image will be stretched and the pixel mapping routine applied.
The image will be compared to other ESO images of the Carina Dwarf and carefully scrutinized for artefacts.
A very big thanks to Olivier and Lars Lindberg Christensen.
If anyone is interested apart from Robert, the European Southern Observatory (ESO) through the astronomer Olivier Hainaut has graciously supplied me with raw images of the Carina Dwarf taken with the 2.2 metre MGP/ESO and 4 metre Victor M Blanco
The image will be compared to other ESO images of the Carina Dwarf and carefully scrutinized for artefacts.
Regards
Steven
I guess that they actually are interested, out there in cyberspace, judging from the number of views. But it would be nice to hear from some knowledgable "deep imaging" practitioners (Ken Crawford springs to mind)
I hope that you do succeed in reliably bringing up details at fainter than 30 V magn. per sq. arcsec.
The fact of the matter is that most of the existing surveys for Low Surface Brightness objects and features have actually been very limited in their depth, due to inevitable constraints on Large Telescope time. (a lot of discoveries of L.S.B. objects, such as the ultra-faint dwarf galaxies around the MW, were made with SDSS imaging data, which is not what anyone would call deep)
Even the new generation of (> 10,000 sq. degrees) Sky Surveys, as typified by SkyMapper Southern Sky Survey(http://rsaa.anu.edu.au/research/proj...ern-sky-survey ) and the VISTA Hemisphere Survey ( http://www.vista-vhs.org/ ) and the Large Synoptic Survey Telescope, will require very multiple passes of the same field, taken over many years, to achieve deep images of various fields.
In fact, the new-generation Sky Surveys are specifically not optimized for depth! (bizarre, but true)
Therefore, I think that there is room for amateurs to make a serious contribution in the area of ultra-deep imaging of galaxies and inter-galactic objects.
The ESO image is a total 31 hr exposure taken through various filters.
Whereas we amateurs use LRGB exposures for galaxy images, the setup is somewhat different for the professional.
The colour data for the Carina Dwarf is a composition of V (visual), U(ultraviolet), B (blue) and I (infrared) filters.
Instead of messing around trying to produce a colour image with a system I have absolutely no experience with I decided to stack the V, U, B and I data to make a luminance image.
The processing involved successive iterations of the pixel mapping routine while making sure the histogram was not clipped but keeping the black point as near as possible to the base of the histogram.
The objective is being able to recognize structures which have a surface brightness of around 30 Vmag per square arcsecond.
To put this into perspective this is over 1000 times fainter than the naturally occurring skyglow from the darkest sites on Earth.
It is why it is so important that noise is kept to an absolute minimum.
For me the image looks more like a random starfield, I'm sure Robert with his keen eye for detail might see something different.
Impressive data processing! It is very hard to evaluate it in details on the re-scaled JPG (which loses lots of information wrt the FITS), but here are some comments:
- The noise texture in the background, and the large-scale flatness of the background suggests you have used a sky-subtraction algorithm (which is OK) using a sampling window that is too small (probably 100pix, while you could push it to ~500 or 1000)?
- the overall depth of the image is really good!
- combining the data over a huge wavelength range UBVI can be counter productive, even if your only goal is to go deep. For instance, the individual U images tend to be much shallower (because the camera is less sensitive in U, and bc the stars are fainter in U), so if you average them with the others without a different weight, you actually increase the noise more than the signal. You could consider to use the following weight for each image:
w = F/t
with t = exposure time (i.e. normalize by exposure time) and
F(U) = 0.2; F(B) = 0.4; F(V) = 1; F(R) = 1.5; F(I) = 0.8
An other way would be to pick a star that is not saturated in any of the images and that has neutral colours in our image, and normalize the frames by the (sky-subtracted) flux of that star.
If you have the time/the patience, you could give it a try, and I would not be surprised if the end result is a little deeper.
Note also that for the astronomers, combining all the filters together is usually not useful. It can help to create a catalogue of positions of the starts, and then go back and measure the flux in the single-filter recombinations U, B, V, R, I.
- the stretch function you apply is very good for displaying the image - but for us the only useful one is the original linear pixels. The stretch functions are really just for display. Anything that screws up the linearity of the data is a problem (and actually, we sometimes have problems with the raw data being not perfectly linear).
Cheers
oli
The image qualifies as nice "pretty picture" but doesn't carry much scientific value.
An astronomer would not be able to use the image and calculate the surface brightness of the galaxy due to the non linear stretching of the data.
Interesting however is that while the pixel mapping function is non linear over the entire pixel range (0-65536), it is in fact close to linear in the low range which corresponds to the brightness of the Carina Dwarf.
It is great that an expert on image processing has taken the time out to give a professionals perspective on the subject.
The image qualifies as nice "pretty picture" but doesn't carry much scientific value.
An astronomer would not be able to use the image and calculate the surface brightness of the galaxy due to the non linear stretching of the data.
It is great that an expert on image processing has taken the time out to give a professionals perspective on the subject.
Steven
Very interesting comments by Mr Hainaut.
In my view, both professionals and amateurs achieve excellent results in the detection of ultra-faint objects and/or features, but by very different methods. This is surely an area in which amateurs and professionals should collaborate and share methods, in order to achieve optimal results.
I have noticed, for instance, that despite all the mathematical-physical knowledge applied in professional image processing, the best amateur imagers are better at bringing out ultra-low contrast detail in E and S0 galaxies than the professional astronomers.
As regards your comments on the scientific value, or lack thereof, of images such as yours, I hasten to point out that the mere detection & display of faint objects and/or faint features, can qualify as a discovery, if the object or feature in question is novel.
I really do need to find the time to consider your image in more detail, but I am currently being henpecked regarding mowing the lawn.....
Very interesting comments by Mr Hainaut.
In my view, both professionals and amateurs achieve excellent results in the detection of ultra-faint objects and/or features, but by very different methods. This is surely an area in which amateurs and professionals should collaborate and share methods, in order to achieve optimal results.
In a way Robert this has occurred here.
ESO's interest in how a piddly amateur telescope (my words not theirs) could image the Carina Dwarf has resulted in them acquiring the mathematical details behind the processing.
In return they have supplied the raw data, given a critique including some some valuable tips.
They also have provided access to all their data files.
This is course is available to the general public provided you know who to ask.
Quote:
I have noticed, for instance, that despite all the mathematical-physical knowledge applied in professional image processing, the best amateur imagers are better at bringing out ultra-low contrast detail in E and S0 galaxies than the professional astronomers.
As regards your comments on the scientific value, or lack thereof, of images such as yours, I hasten to point out that the mere detection & display of faint objects and/or faint features, can qualify as a discovery, if the object or feature in question is novel.
My "pretty picture" comment was in the context of the data supplied by ESO. Since ESO included infrared data for the Carina Dwarf, it seems that photometry was the major objective.
Stacking their images and doing a non linear stretch is not exactly in the spirit of photometry and was probably reflected in Olivier's E-mail.
That's not to say that Rolf's ultra deep sky imaging or even the my own version of "mathematical jiggery pokery" is without scientific merit when applied to activities such as sky surveys.
Quote:
I really do need to find the time to consider your image in more detail, but I am currently being henpecked regarding mowing the lawn.....
And with the upcoming fire season I need to start cutting the grass on my 10 acre property.
