Nine Panel Mosaic of the VSNR in NII

[avandonk]

This is over 50 hours of 32 min exposures in 3nm NII of the Vela SNR.

FoV 9x8 degrees. 52MB


http://d1355990.i49.quadrahosting.com.au/2017_04/VSNR_NII_9P_.jpg


No noise reduction or any other enhancements.

Exposure details 32 min each sub using CCD flushing and the slowest download speed. This enhances signal to noise for the really dim stuff.

I have all the OIII data as well.

Bert

[RickS]

That's a very impressive FOV, Bert. Some wobbly bits in the joins between panels but you can easily overlook them at such a grand scale.

Cheers,
Rick.

[Atmos]

That is properly huge Bert!! Very nice result!
Going for a Bi or Tri colour? Going to be a fascinating image.

[gregbradley]

Sensational Bert. One of your finest. Its amazing that one star blowing up could make such a mess of stuff.
It must have been an enormous star.

Greg.

[Andy01]

Crikey Bert, you didn't leave anything on the table there did you!
That is EPIC in scale. :eyepop:

As I'm learning about mosaics myself, may I ask what software you're using to register & merge the images?

Cheers
Andy

[clive milne]

Now that is special Bert...
I suspect it would a good subject for continuum subtraction.

[astroron]

Splendiferous there Bert:eyepop:
It is amazing how small the Pencil Nebula (NGC 2736)
is in the whole scheme of things.
Wonderful image in full resolution.
Cheers:thumbsup:

[SimmoW]

THE BERT IS BAAAACK!

Bloody amazing, deep and huge image, well done. Probably the largest mosaic of the area I've seen.

[strongmanmike]

Well well well, if it ain't the old mad scientist himself, welcome back Bert :hi:Awesome and rather laaarge field :eyepop:...you've had a few cracks at this region if I recall..? Can understand why, that's for sure. I'll have an FSQ106 soon so I am looking forward to having a crack at a bit of this myself :D

One thing though :question:...how come you have a picture of Rafael Nadal as your avatar picture..?

Mike

[Bassnut]

Gee, thats big Bert, excellent. And the perfect reference for so many narrowfield opportunities.

[avandonk]

Andy I am using Pixinsight to register and merge the images at 32 bits (floating point) and linear i.e. unstretched.
It is far easier to match panels and feather them mathematically when the data is linear.

At 16 bits all the 3nm narrow band nebulae data is below the first nine bits or 512 levels in 65,536 even with a 32 minute exposure. The really faint stuff is in the first few bits. In a 32 bit image the nebulae data is spread over the first 18 bits or 262,144 levels. By the way it is important to use the 32 bit xisf format for dark flat correction and stacking etc.

Registar is much better at aligning images than Pixinsight because it distorts locally to account for the fact that any image is a flat projection of a spherical real image. The wider the FoV the more noticeable this is.

There is a way to get Pixinsight to distort more locally. Set up the Star Alignment window as below. Setting Spline Smoothness to zero does this. Note unlike Registar, Pixinsight keeps the secondary image borders rectilinear.

I then use GradientMergeMosaic to merge the output of Star Alignment.
GradientMergeMosaic does a very good job of producing seamless mosaics. I set the feather radius to 60 pixels for my data. Sometimes a bright star near the joins can affect the much lower background. Either trim a bit off the image or change the feather radius.

Hope this helps.

Bert

[avandonk]

Thanks Bassnut. My aim is to try to image the faint stuff most others ignore due to time constraints.

With my wide fast field it can give others with longer optics a guide to where to image any resulting interesting bits in better detail.

Each of these panels is at least 11x32 min or six hours. To get the same signal to noise at F5 would take at least 20 hours.

Bert

[avandonk]

Thanks all others for any comments.

Mike that is a picture of me at 24 years of age when I worked at Kodak in 1973 when I was young and foolish.

I have deteriorated a bit since then. Gotten greyer and a bit grumpier or is that old and still foolish.

When you read in someone's Obituary that they did not suffer fools. It really means they were a pain in the A to all others.

A bit more detail about the imaging protocol. I use CCD flushing for all exposures and the slowest download speed. This really helps to keep the noise down. All darks are also done exactly the same way. I use one hundred Bias frames and at least sixty Darks.

At least thirty flats for the flat correction with correction for darks and bias. Another good thing about Astrodon Filters is that I use a flat taken with the NII filter to do the flat correction for all the other filters. Even with my F3 optic the narrow band filters show exactly the same relative recorded profile as the RGB and L.

I took meticulous care when installing the filters into the CFW to not introduce any dust. Even now I only have a few dust motes on the rear surface of the corrector lenses in the rear of the RH200. None on the filters.

With a lot of trial and error I have found the best conditions for integration see below.

When I have more than twenty subs I use 2.0 for Sigma Low.

Bert

[avandonk]

Thanks Greg for the encouragement. The only thing that really works is practice, practice and yet more practice. There is always room for improvement. This will keep me occupied until I croak and rejoin the dust from where I came.

Greg nearly all the elements apart from Hydrogen and Helium that make you and all around you came from stars such as this.

I was from many messes like this that our Solar System formed.

I am sure you knew this.

It has taken 13.7 billion years to get to where we are today on our tiny Space Ship Earth. I first heard this from Carl Sagan many years ago.

Bert

[strongmanmike]

Yes...well...we all get old...coincidentally, my wife was cleaning out cupboards last weekend and found this photo of me at 24 (http://www.pbase.com/strongmanmike2002/image/165240389/original)in a paper bag of old prints tucked away in a corner of a box....:sad:

:lol:

[SimmoW]

Hey thanks for the PI tips Bert, I've done a couple of mosaics but am yet to process them in my backlog. Love this image, yes I already said it but it's worth repeating!

[RickS]

What sort of rejection percentages do you get with those parameters, Bert?

Cheers,
Rick.

[gvanhau]

Nice Mosaic Bert; Much better than one I saw from you one or two years ago. Congrats.

[DJT]

Wow! Didn't realise quite how big this area was. A great image, congrats!

[cazza132]

Dude, that is some hard yards - 50 hours! Dedication! Rewards are there to see - awesome image in the making! Interesting choice of filter - not one that that I have seen used too often. Looking forward to seeing the final result!

[alocky]

Hi Bert - as the others have said, it's rare to see the whole of this remnant at once, and the huge effort in data acquisition and processing to produce this is staggering!
One question on your integration in PI, do you tune the rejection parameters for each dataset? I often find that the optimal thresholds vary considerably between objects, and I keep adjusting the sliders until I start to see signal getting rejected, then back them off. I appreciate this may cause problems in a mosaic though, have you found this?
Cheers
Andrew

[avandonk]

Andrew ideally when doing mosaics one would prefer that the Moon was not there, the seeing and transparency were the same, each panel had the same number of in focus high quality subs, and a lot more.

Over many nights and different weather conditions this is a forlorn assumption.

When we have so many variables it is foolish to introduce more variation in the data by fiddling with the processing parameters. This only leads to problems matching panels for the mosaic.

All data sets for each panel are processed identically for this reason.

There are far more problems with the OIII data as this is more susceptible to atmospheric attenuation and light pollution even with 3nm data. Atmospheric pollution and dust and don't even mention the Moon, has far more effect on OIII than NII, HA or SII.


I have all the data for OIII for the same area. Unfortunately a few too many of the panels were ruined by atmospheric conditions. The resulting mosaic was not good enough to produce a colour image with the same quality as the NII data.

Bert