FWHM Aperture tool

[Jack Harvey]

How about a cursor based aperture tool to allow measurement of FWHM once the image scale has been put in?

[Carlos Milovic]

This has been on the debugging table for a long time my fault Actually, right now it occupies the fourth place in my to do list surelly will be ready this year.


BTW, this is my to do short list:
1.- Fourier Transformations (from and to fourier space)
2.- MultivariateCurves (something like the Hue/Saturation tool)
3.- AutoSTF (a quick/dirt way to set the STF with statistical data)
4.- ReadPSF (dynamic Gaussian PSF extractor)
5.- HDRC (a HDR algorithm)

[Niall Saunders]

Hi all,

If we are heading in the direction of FWHM analysis (hurrah  ), then - because we already have code in place to identify stars, and because we already have the 3-D intensity plot module, would it not be just one step forward to create an analysis tool that shows FWHM 'across the image' - allowing users to determine whether their CCD sensor is truly orthogonal to the incoming light cone, and whether they have optical abberations in any areas?

This might be more useful than simply providing FWHM for an individually selected star.

Just a thought . . .

[Carlos Milovic]

Sounds nice But one step after another

[Niall Saunders]

Pah,

Such little steps Carlos  You would have been of no use climbing down the ladder of "the Eagle" back in '69 

However, little steps DO have a tendency to help you avoid falling on your backside !!

Can I ask, therefore, that the FWHM analysis tool at least provides :-

X-axis FWHM
Y-axis FWHM
'Diagonal' FWHM (based on d = sqrt(x^2 + y^2) )
and the 'angle of rotation of the solved diagonal'

all of which are simple calculations once the ability to detect and calculate FWHM has been implemented.

I am sure that you probably already have this in mind, but hoped you wouldn't mind me asking anyway.

Thanks,

[Carlos Milovic]

I wrote a more sophisticated code, but still has some bugs to deal with. It calculates both the mayor and minor axis of the elipsoid, and fits a general anisotropic gaussian function. Then you get both fwhm and the rotation angle (along with the center, intensity, etc etc etc).

[Niall Saunders]

Whoa, Carlos

Big steps already 

At least we were on the same wavelength although, as usual, you are several harmonics 'up' from me (what with your anisotropic ellipsoids, etc  )

Looking forward to the results.

[Jack Harvey]

Cool

[mmirot]

We need this tool too.

[Andres.Pozo]

I like the Half Flux Diameter (HFD) measure because it is VERY easy to implement (the algorithm is very similar to the centroid algorithm), it is fast and the result is more stable in bad seeing conditions. I have used it for focusing and the results IMVHO were better than with FWHM.

http://www005.upp.so-net.ne.jp/k_miyash/occ02/halffluxdiameter/halffluxdiameter_en.html

[Niall Saunders]

Hi Juan,

I am 'replying' to this message - just to keep the topic alive.

Right now I feel that I need more 'inspection tools' for PixInsight.

I want to be able to evaluate FWHM, preferably on a star-by-star basis, and would LOVE to be able to create a DBE-Background type of image from the FWHM array information. Then I can look at the image as a 3-D plot and see what the optical behaviour of my setup actually is.

Now, MOST of this could be implemented as a PJSR, especially if the FWHM was available as a PCL call - and if all the other PCL core processes returned useable properties (for example, a user could call a 'StarDetection' core PCL process that would return an array of information containing X and Y coordinates of the centroid, FWHM in the X and Y axes, peak ADU at the centroid, etc.), which would mean that 'somebody' could themn tackle the PJSR (to save Juan having to implement it at PCL level )

Any takers?

Cheers,

[Carlos Milovic]

Hey Niall

I wrote the StarStatistics class, that is supposed to be incorporated on the PCL as soon as Juan feels it is good enough
It wasn't a very robust algorithm for centroid searches (you needed to be fairly close, and have an aperture small enough to not include too much sky). Anyway, for a FWHM value, I think it was functional. Let's see if we retake this for 1.6.1 or other release in the 1.6.x series.

[Niall Saunders]

Thanks Carlos,

I think we need to get 'something' out into the public domain. Once it is there, it could be tweaked to suit.

Cheers,

[Carlos Milovic]

It's fine for me. After all, its not big secret I use a minimum square approach.
I'll look for the code once at home, and release it under GPL, so you all may contribute.

Oh, almost forgot that I should review it, and comment the steps