PixInsight Forum (historical)

PixInsight => General => Off-topic => Topic started by: RobF2 on 2011 February 06 02:18:25

Title: Morphological transform correct tool for softening stars?
Post by: RobF2 on 2011 February 06 02:18:25

I've been reading up on the Morphological Transform tool, but would like to check I'm headed in the right direction.  My Newtonian tends to burn out the brightest stars and leave "flat tops" on an intensity plot for the less bright ones.  Is MT the right tool to help soften them to a more Gaussian appearance, or should I be looking at using another process?

I really haven't bothered with separate star processing layers in the past, but curious for any recommendations.

thanks,
Rob

Title: Re: Morphological transform correct tool for softening stars?
Post by: Simon Hicks on 2011 February 06 04:40:17

Yup, this is the right track...though maybe others have different routes.

Create a star mask that selects just the brightest stars, i.e. the saturated ones. Let the star mask expand a bit. Then use MorphologicalSelection, maybe 6 iterations, Amount = 0.2, Selection = 0.2, Structuring Elements size = 5, Circular. Use this as a starting point and play with the parameters to see what works for you. I find it softens the edges, shrinks the star and gives more of a rounded top rather than the saturated flat top. If you have any colour in the dimmer parts of the star then you will see this progress towards the core as well.

Time to play  ;)

Title: Re: Morphological transform correct tool for softening stars?
Post by: Nigel Ball on 2011 February 06 15:32:56

I'd suggest also trying Size 3(9 elements) and see which you prefer

Title: Re: Morphological transform correct tool for softening stars?
Post by: RobF2 on 2011 February 07 03:22:29

Thanks Simon and Nigel

I could have used your notes yesterday Simon.  I spent hours playing with this thing and reading back through forum threads.  It does seem like the selection slider is the way to go, and I agree totally that 3x3 or 5x5 elements seem to give a good response with 6 or 7 iterations.  I think what you've described would save most mere mortals a LOT of fiddling to get a passable output.

2 comments after all this:

1.   This thing is a BEAST!   >:D   Its incredibly and deceptively powerful - I gather "ways" are a means for layering different dimensions of structural elements?  Woooh dude..... O0
2.   I don't like to whinge too much about documentation normally, BUT, this process appears to have almost none of the usually very helpful auto prompts appearing - couldn't we at least have some of those until a proper documentation page is created - please...... :'(


For others passing this way, perhaps also read up on:
http://pixinsight.com/forum/index.php?topic=500.15 (http://pixinsight.com/forum/index.php?topic=500.15)
http://pixinsight.com/forum/index.php?topic=2381.0 (http://pixinsight.com/forum/index.php?topic=2381.0)
http://pixinsight.com/forum/index.php?topic=2717.0 (http://pixinsight.com/forum/index.php?topic=2717.0)