Changing FWHM on Image Integration

[pvelez]

I've noticed of late that many of my RGB images have a green or aqua halo around stars, particularly brighter stars.

Applying the FWHM Eccentricity Script to the Masters, I see that the FWHM of the files differ - with G the largest and R the smallest. Now the odd part of all this is that the subs that I integrated had a reasonably consistent FWHM across all filters - as assessed with the Subframe Selector script.  Perhaps I am misunderstanding the FWHM parameter. However I am surprised that I have bloating in G and to a lesser extent in B using the same integration settings.

Am I missing something?

Pete

[msmythers]

Pete

Do you have local LP? That and high haze can cause this. I have to blink all my subs because I live in a year round, very hazy coastal area. Here is an example from blink with a wide field shot.
https://drive.google.com/file/d/0B13r3kYqTf8NajRFR1VPTVVZT2c/view?usp=sharing You can see Rigel and the other stars change as the haze moves. This was over a 36 minute period.


Mike

[pvelez]

Mike

I'm imaging from the iTelescope facility which is on the SSO site at Coonabarabran in NSW, Australia. Its a research site and very dark. So its not LP. I agree that there might be some high cloud or haze about - but I doubt that is contributing to the issue.

I have a suspicion it is a feature of the Image Integration process. To give you a concrete example, - I just ran the Subframe Selector script on 29 frames taken in G - the highest FWHM I accepted for the Master was 3.298. I then combined them with Image Integration using Winsorized Clipping. Checking the resulting Master with FWHM Eccentricity script, I see that the median FWHM is 4.321. So for some reason in the integration process, my FWHM has jumped up by 1 pixel.

Ahhhh - just twigged - rookie error. I had Subframe Selector displaying in arc seconds not pixels. Change that and it all makes sense. My maximum FWHM in the subs was actually 4.71 pixels. The Image Integration process brought that down to 4.321. D'oh! My bad.

I'll leave this post as I've typed it as a warning to others to check the units!

Pete

[msmythers]

These things happen. 8)


Mike

[pvelez]

They certainly do Mike

A Homer Simpson emoticon is called for here!

Pete