Vela SNR Mosaic in 3nm NII and OIII

[avandonk]

This is just the start of what will be a 4x4 mosaic in NB and RGB. Even if it takes some years because of weather.

Full res image FoV 6.0 x 3.2 degrees. 12MB

http://d1355990.i49.quadrahosting.com.au/2014_01/VSNR_NB_NOO.jpg


Exposures were 20+ X 16 min for OIII and NII for both panels.

I am going for the maximum practical faint detail my system will record.

Bert

[strongmanmike]

Pretty impressive field that Bert :thumbsup:

Mike

[gregbradley]

Very nice Bert. One of my favourite objects. Lots of details.

Greg.

[RickS]

Should be a stunning mosaic, Bert. The data looks to be fantastic.

[batema]

Amazing photo.

[avandonk]

Thanks for the comments. I have been up all night now for three nights in four. What makes it worth it when you see something you've never seen before.


I tend to stack my data 1.5 times the native pixel count of my sensor. This with dithering and lots of frames increases the resolution due to the under sampling. One thing this allows me to do is then bin x2 or 50% which increases signal to noise by a factor of four. The image is now 75% of native pixel count about right for mosaics and the stars are now perfectly round.

This image is a starless version derived from binned x2 data and upsized to native sensor size. 9MB

http://d1355990.i49.quadrahosting.com.au/2014_01/VSNR_NB_NOO_nostars_.jpg

Bert

[astroron]

Incredible image Bert. :eyepop:
I always look at your images with awe.:)
Cheers:thumbsup:

[avandonk]

Here is an image at native pixel size with stars derived from the x2 binned data. 11MB

http://d1355990.i49.quadrahosting.com.au/2014_01/VSNR_NB_B_N.jpg


The stars are a little bit bigger but now perfectly round compared to the image from the first post and the faint nebulosity is now far better defined.


Bert

[RickS]

Have you looked at the Drizzle algorithm at all, Bert? Hopefully it will be implemented in PI before too long.

Binning x2 will increase signal by 4 times and shot noise by sqrt(4) times for a net improvement in SNR of 2 times. Suffering from too many nights at the scope? :)

Cheers,
Rick.

[avandonk]

What shot noise? Do you mean Poisson noise?


Bert

[RickS]

Yes, shot noise aka photon noise aka Poisson noise: http://en.wikipedia.org/wiki/Shot_noise

[avandonk]

I am well aware of all sources of noise. I do have a degree in Physics.

Do you think that nutters such as myself would buy fast optics just to overcome these limitations of signal to noise if they did not work?

The only way to overcome or minimise Shot or Poisson noise is a fast optic where the stream of photons are like a river rather than a noisy trickle!

Drizzling is just upsizing and stacking!


Bert

[cometcatcher]

What an amazing, delicate looking cobweb of gasses. Suns definitely look prettier when they blow up. :)

[RickS]

My comment on Drizzle applied to this:



It seems like Drizzle may be a better way of doing the same thing, i.e. wringing some additional resolution from dithered data. It's a little more complicated than just upsizing and stacking.

My comment on shot noise and SNR applied to this:



I was pointing out that you have increased your signal by a factor of four but not your SNR which has only increased by a factor of two. I was also kind enough to assume that you already knew this and had just made a silly mistake.

Cheers,
Rick.

[avandonk]

It is a pointless exercise to argue about the intricacies of signal to noise here.

Let us just say that I do not need or want a lecture in statistics. If you want to be pedantic go to a more suitable forum or is that fora?

To tell the truth all I care about is a clearer image of heavenly bodies so my natural urges can be satisfied by a certain frisson of manipulation of my telescope.

Bert

[David Fitz-Henr]

I believe you are correct here Rick; whilst upsizing and stacking can result in smoother data and some marginal improvement in resolution, the drizzling process is not the same and can produce much larger gains in resolution for undersampled images. There is a lot of information on the web about drizzling - here is a good site that explains the concept quite well with a worked example: http://www.asiaa.sinica.edu.tw/~whwang/gallery/random_notes/drizzle/index.htm

Bert, what is the image scale of your system (arcsecs / pixel)?

[Bassnut]

Spectacular Bert, interesting you went with Nii

[Star Catcher]

Fabulous capture and processing to show the structures.
Ted

[DavidNg]

Wow, look really impressive already.

[avandonk]

My image scale is 3.08" per pixel.

I do understand information theory. Farting around with a single image is meaningless.

I dither my images by many pixels. This means I am sampling the same putative image many times.

I typically collect twenty plus images.

Bert

[graham.hobart]

That is a beautiful picture
Thanks for showing it
Cheers
Graz:thumbsup:

[David Fitz-Henr]

Well, the web reference is useful to explain the main concept of drizzling for those that may not be familiar with how it works. I believe it(drizzling) was used on the Hubble deep field images to regain much of the resolution that is lost due to the undersampling by the WFPC2 camera.
As you imply though, our actual images won't replicate this exact outcome; I believe due mainly to random dithering, high freq noise, limited sub frames available, etc. I'm also not clear to what extent the SNR would suffer (per output pixel) since (as I understand it) drizzling is a process that "constructs" a higher resolution image from multiple (lower res) subframes, whereas with simple stacking of n subframes (using say the mean on the original pixels) the SNR is increased by sqrt(n). I assume that for sufficient sub frames there will also be some "stacking" of the drizzled image output pixels which will increase the SNR at the higher res pixel size?
So ... given that your system is undersampled it would be very interesting to compare upsized/stacked vs drizzled images from the same set of sub frames if you ever have the opportunity. I believe that drizzle is being incorporated into some of the popular software and may also be available standalone.

[allan gould]

Great image, Bert

[avandonk]

What I do is collect many images at 3.08" per pixel. I always use the RBI function on the PL16803 camera. This gives far less noise due to residual signal due to brighter stars when dithering.

By experiment I have found out how many images are needed for RGB and NB beyond which is there is not much more gain in both signal to noise and resolution enhancement. Of course if you REALLY need to get a bit extra S/N and resolution, more images are always better. If one collects data till the end of time the image will still not be perfect!

I correct the frames for darks and flats. I use at least thirty darks and flats to make the master darks and flats. I even correct for darks with the flat frames. This ensures that a minimum of noise is injected into the data due to the correction process.

I then upsize the images by a factor of 1.5. This now gives me about 2" per pixel.
Note this is the resolution of the width of the pixel not the corner to corner, which is root two this resolution. I could go into information or sampling theory here but I won't.

By now stacking these upsized frames, I get an image frame where both signal to noise and resolution are enhanced.

Drizzling is quite a bit different to this as they have far less data frames so even a more selective mathematically valid sampling is required. They have the added variable of sensor orientation in the case of Hubble.

It is pointless stacking images without dither as any residual noise is enhanced along with the signal.

My method is just elegant brute force. What essentially is happening is that I collect lots of real signal while ensuring that the noise is suppressed.

I can then afford to increase resolution at the expense of more noise or conversely decrease resolution by binning the upsized image to get better signal to noise. I estimate this to be about a factor of four. It is most probably a bit less due to the Poisson or shot noise. Dithering and many frames lowers the inherent unavoidable shot noise of the very weak signals we are attempting to image.

Finally fast optics means that you collect data far faster than any noise no matter its source! The ratio of the square root of N photons over the N photons goes down as N increases. This is Poisson or shot noise.

Sky glow and light pollution are the exceptions as these are both collected just as fast.


Bert

[Suzy]

:lol: Agreed!

Fabulous image Bert! :eyepop:
Posted on the IIS fb page. Thank you for sharing it with us. :)

[g__day]

Simply wow! Took my breath away!

[avandonk]

Suzy all my images are put up at full resolution for a very good reason. They look better that way!

All my images on IIS can be used by anyone as long as it is not for profit.

If for profit then a simple acknowledgement is all that is generally needed.

Bert

[Suzy]

Thank you Bert, yes, I know this is a very generous thing you do as I've asked in times gone by if I can print off photos of your glorious work and you've been more than happy for me to do so. Thank you for sharing the universe with us so freely Bert. It's great that a good amount of people still do this and understand what it really means to us (as you know, some won't with their copyright work!). I won't share anything on the IIS fb page unless I can credit the person and/or provide a link-it's the right thing to do. I just love your work Bert- have done so for many years now. Thank you! :)

[Harb]

Yes I just had another look at the full res image.....it really is breathtaking work........if one day I can ever produce that sort of work I will be pretty happy.......must be very satisfying.

[avandonk]

Fred there is some conjecture whether at F3 that 3nm NB filters have 'leakage' from HA to NII and the converse due to frequency shift caused by the modulation of the incoming beams dependant on the angle at which they are incident on the 3nm filter.

I have found that NII data shows more detail in nebulosity especially the faint stuff.

The only way to solve this conundrum is to image a pure HA or NII object with both filters.

Bert

[avandonk]

It is far deeper than that Suzy. Long ago many of the atoms that make you or me were made and dispersed by super novae. You are seeing my pathetic attempt to image our real ancestors.

Our Solar System originated by the gravitational aggregation of just these atoms. This needs a very cool place for the kinetic energy of these particles to get to a low level so that gravity wins.

Nuclear synthesis in the heart of stars can only go as far as Iron. All of the heavier elements are produced in super novae.

'We are indeed children of the stars.' Carl Sagan.

[David Fitz-Henr]

Yep, I also dither all of my images and usually aim to collect around 20 luminance sub-frames for a given target. I also realise that the Hubble images are quite specifically dithered in terms of offset and angle. My interest though is in the practical application of drizzling to amateur images; to what extent is random dithering compensated for by taking additional sub-frames, etc. Ie. will the purported gains in resolution for undersampled amateur images be realised?



You know, I think Carl may be right; the more I stare at your great image of the Vela SNR the more certain I am that I can see a family resemblance ... :rofl:

[avandonk]

David resolution gains can only be made if the optic has better resolution than the sensor.

First law of the Universe 'there is no free lunch'.

By upsizing an image it is just being resampled with a loss of overall signal to noise. With many of these randomly dithered upsized images then stacked, both S/N and resolution are enhanced. More images then more enhancement.

Have a look here

http://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_th eorem


Shannon was on to this problem when I was just a young boy.

His Theorem applies to temporal resolution of a time dependant signal. It is equally valid for spatial resolution of a two dimensional array.

Bert

[RickS]

Bert: can you explain how resolution is enhanced by this process? I can see how it could produce a smoother looking result by interpolating followed by averaging but I don't understand how it could produce a real gain in resolution. If I'm missing something please enlighten me.

Cheers,
Rick.

[avandonk]

Rick when a single image is upsized it is as if it was taken with a sensor with smaller pixels.

When many dithered upsized images are then stacked. The inherent resolution of the optic is sampled at different positions. This is where the higher resolution spatial information comes from.

This only works if the optic has better resolution than the sensor.

With enough frames by upsizing by a factor of 1.5 my 4096x4096 pixel sensor is now effectively a 6144x6144 pixel sensor.

This does not come for free as noise increases.

To put it into some sort of perspective my sensor is 3.08" per pixel. The resolution of a single image is at best 2x3.08" or 6". By stacking at native pixel size this will improve slightly.

By upsizing by a factor of 1.5 we now have 2" per pixel and the resolution of a single image is 4". By stacking at an upsized size we get just a bit better than 4" resolution. We need to take into account seeing and tracking which can be as bad as 2" and is additive.

So can you see my cunning plan? I do not need to worry much about seeing like all the other astrophotographers at longer focal lengths. The PMX mount tracks better than 1" or far less than one sensor pixel.

In fact there is a very high end studio digital camera ($50k+) that moves the sensor relative to the optic axis by about a pixel and takes three exposures in very rapid succession at slightly different positions and this improves resolution.

Bert