Star masked Bubble, your thoughts?

[ollypenrice]

I've been beavering away trying to find ways of keeping stars smaller in widefield images, notably Ha RGB ones. I'm currently favouring a variant of the method proposed by Doug German.

However in my version I make a starmask from the linear image and leave the white point where it is on the star mask. Moving it to the right as Doug does brought nebulosity onto the mask for me. I do one or two stretches using this mask and then I flatten the image and use the partly stretched one to create a new mask on which the stars are now a little larger. This avoids them being only partly masked in the later stretches. I then stretch again once or twice.

Very large stars (maybe only one to four in an image) defeat this process. For these I take the linear image and make a star curve which flattens early. This gives a normal background but smaller stars. Before it is fully stretched I paste this onto the main image and put big feathered selections around the stars in question, delete what is outside the selection and then play with levels, curves and colour balance to get the small top layer stars to blend in seamlessly with the main image.

Lastly it had 1.5 iterations of Noel's 'Make Stars smaller'. All this mucking about with stars can affect their shape but if you select the them in a tight feathered selection (eg using Noel's 'Select Brighter Stars' a slight blur seems to restore them.)

Anyway here's my test case, the Bubble, which needs an almighty stretch to get the faint signal into view but which is also stuffed to the brim with stars, wrecking my earlier versions.

Be honest about what you think of it. I've been gawping at it for so long that I've lost the plot...

TEC140/Atik 4000 mono, HaRGB. I think the TEC lens did me proud regarding detail inside the Bubble.

Olly

Biggest; http://ollypenrice.s...Q37VN3&lb=1&s=A

[davew]

If you want honesty I'll give it !

In the main it's worked ok. Star colour is holding up well. At a reasonable distance all seems to be tight but I downloaded it and can see dark rings around many of the mid to small sized stars. The larger ones have reacted more like you wanted.

The general colour balance is nice to my eyes and the background particularly good. The dust is showing nicely and is that a rather nice reflection top right ?

If you didn't really want honesty then it's the best Bubble I've ever seen

Dave.

PS. I think Doug does it the way he does because it's aimed at noisy DSLR images.

[Tim]

I'll have a closer look when I get home from work Olly and read your post properly.

I think I read about a filter the other day which gets rid of stars, that you can use in conjunction with narrowband filters to just have the nebula. Might have dreamed that though?

Tim

[ollypenrice]

If you want honesty I'll give it !

In the main it's worked ok. Star colour is holding up well. At a reasonable distance all seems to be tight but I downloaded it and can see dark rings around many of the mid to small sized stars. The larger ones have reacted more like you wanted.

The general colour balance is nice to my eyes and the background particularly good. The dust is showing nicely and is that a rather nice reflection top right ?

If you didn't really want honesty then it's the best Bubble I've ever seen

Dave.

PS. I think Doug does it the way he does because it's aimed at noisy DSLR images.

Great feedback, Dave, and thanks. You are quite right about the stars. Truth is I fixed the dark haloes using Nik's trick of clone stamp set to Lighten for those stars around the interesting bit and thought I'd get round the rest some other time. But you rumbled me! It would be good to fix that issue generically in the workflow. The TEC is mainly a visually corrected apo so it isn't as tight as a drum in CCD blue. I suspect this is the origin of the dark haloes but when an image is as Ha dominated as this it's hard to avoid them.

Olly

[davew]

Blue was sent to give pain to imagers everywhere. I've seen planetary imagers close to tears !

Your / Nik's technique works where you've done it. Missed them no matter how close I looked.

The labour of love you're showing now will become hate if you need to rescue dozens of stars every time.

From what little I know the depper red will also do some damage.

Dave.

Edit - Looked closer still in PS. Red is causing problems and blue makes it look worse. I think.

[ollypenrice]

Heh heh, I know!

Another way to reduce dark haloes quickly is to duplicate the image, lift the bottom of the curve on the lower layer sharply then flatten it off. Then you can whizz round with a soft edged small eraser, dabbing it on the offending stars.

Olly

[Tim]

I have always wondered if doing a new star mask for every stretch would be useful. The star mask routine in Pixinsight is very adjustable, if only it could more easily work like layer blending in PS it would be much easier. I know you CAN do it with the pixelmath function but it's not as intuitive as the opacity sliders in PS.

For big bright stars I rather like Ken Crawfords method of using the pinch filter to shrink them and then paste them over the top, which keeps star shape and the surrounding background colour too.

Do you use any deconvolution routines before you start processing Olly?

[Gina]

That's a lovely image

[ollypenrice]

I have always wondered if doing a new star mask for every stretch would be useful. The star mask routine in Pixinsight is very adjustable, if only it could more easily work like layer blending in PS it would be much easier. I know you CAN do it with the pixelmath function but it's not as intuitive as the opacity sliders in PS.

For big bright stars I rather like Ken Crawfords method of using the pinch filter to shrink them and then paste them over the top, which keeps star shape and the surrounding background colour too.

Do you use any deconvolution routines before you start processing Olly?

I've never got any sense out of deconvolution, Tim, and never use it. Maybe if I understood it...??

Olly

[Tim]

Well I can't pretend to understand it all myself, but essentially it is a process of mathematically reducing the unwanted effects of distortion introduced to an image, either through the optics, or perhaps atmospherics.

Essentially it is a sharpening tool, but it does this by tightening everything up, and can be applied via star masks just to stars, just to the DSO, be it galaxy or nebula, or both. It makes a big difference to images of galaxies with plenty of dust in for instance.

Here's a little light reading material

http://www-int.stsci.edu/~mutchler/documents/ACS_APEX_NIST.pdf

Cheers

Tim

[Tim]

Oh, and you know what I'm going to say, Pixinsight has the best implementation of this tool that I have seen so far, along with a preview script to try out different settings.

[ollypenrice]

Well I can't pretend to understand it all myself, but essentially it is a process of mathematically reducing the unwanted effects of distortion introduced to an image, either through the optics, or perhaps atmospherics.

Essentially it is a sharpening tool, but it does this by tightening everything up, and can be applied via star masks just to stars, just to the DSO, be it galaxy or nebula, or both. It makes a big difference to images of galaxies with plenty of dust in for instance.

Here's a little light reading material

http://www-int.stsci...S_APEX_NIST.pdf

Cheers

Tim

Light as depleted uranium! I'll do my best!!

I rarely sharpen stars myself. In fact when under sampled, as with the Tak and reducer, I sometimes blur them a little.

Olly

[Tim]

From Wiki:

Optics and other imaging

In optics and imaging, the term "deconvolution" is specifically used to refer to the process of reversing the optical distortion that takes place in an optical microscope, electron microscope, telescope, or other imaging instrument, thus creating clearer images. It is usually done in the digital domain by a software algorithm, as part of a suite of microscope image processing techniques. Deconvolution is also practical to sharpen images that suffer from fast motion or jiggles during capturing. Early Hubble Space Telescope images were distorted by a flawed mirror and could be sharpened by deconvolution.

The usual method is to assume that the optical path through the instrument is optically perfect, convolved with a point spread function (PSF), that is, a mathematical function that describes the distortion in terms of the pathway a theoretical point source of light (or other waves) takes through the instrument.^[3] Usually, such a point source contributes a small area of fuzziness to the final image. If this function can be determined, it is then a matter of computing its inverse or complementary function, and convolving the acquired image with that. The result is the original, undistorted image.

In practice, finding the true PSF is impossible, and usually an approximation of it is used, theoretically calculated^[4] or based on some experimental estimation by using known probes. Real optics may also have different PSFs at different focal and spatial locations, and the PSF may be non-linear. The accuracy of the approximation of the PSF will dictate the final result. Different algorithms can be employed to give better results, at the price of being more computationally intensive. Since the original convolution discards data, some algorithms use additional data acquired at nearby focal points to make up some of the lost information. Regularization in iterative algorithms (as in expectation-maximization algorithms) can be applied to avoid unrealistic solutions.

When the PSF is unknown, it may be possible to deduce it by systematically trying different possible PSFs and assessing whether the image has improved. This procedure is called blind deconvolution.^[3]Blind deconvolution is a well-established image restoration technique in astronomy, where the point nature of the objects photographed exposes the PSF thus making it more feasible. It is also used influorescence microscopy for image restoration, and in fluorescence spectral imaging for spectral separation of multiple unknown fluorophores. The most common iterative algorithm for the purpose is theRichardson–Lucy deconvolution algorithm; the Wiener deconvolution (and approximations) are the most common non-iterative algorithms.

[Tim]

Notice the bit about early Hubble images being 'fixed' by deconvolution.

[Tim]

Hmm, all this is making me want to go and process the backlog of image data I've been gathering!!!!!

[Tim]

Which includes the bubble at 2800mm

[RikM]

Deconvolution seems to work on some images but is hopeless on others. I find CCDsharp works quite well and it allows you to set a small preview box and run the routine on that a few times to get the settings right before applying to the whole image. Where I am, all the heat rising from nearby rooftops gives generally quite poor seeing and deconvolution seems to help tighten it up.

[ollypenrice]

Which includes the bubble at 2800mm

Wowzer! Remember Rob's immortal Bubble in his 14 inch?

Olly

[Tim]

Wowzer! Remember Rob's immortal Bubble in his 14 inch?

Olly

Yep, that was one of the images which prompted me to get a bigger scope, I have it saved in my special astro folder, which is full of HST pics, APODs, and forum favourites

I've actually got the Ha and Oiii and could do a bicolour image really, but I was hoping for some Sii, forgot about it though when M1 came into view...