Diffraction Spike Removal?

[dnault42]

Hi guys,

Is there a way to remove diffraction spikes from an image?  I'm fairly new to PixInsight and I apologize in advance if this has been asked before, but I certainly couldn't find any mention of it (I did find one thread talking about adding them).  It seems like this should be possible with deconvultion, but at the cost of a loss in signal-to-noise ratio in the area covered by the spike.  I'm considering moving to a Ritchey-Chretien telescope, however I'm not a fan of the diffraction spike effect the way others are and would only make that choice if I knew I could at least mitigate the aberrations before hand.

Regards,
David

[Nocturnal]

Hi,

those spikes blow away the underlying data so it's hard to make something out of nothing. There are essentially two methods:

- copy neighboring data onto the spikes (this can lead to strange effects if you copy stars or other bright objects)
- take two or more sets of data with the camera rotated and do a sigma combine to reduce the spikes before post processing. This works for bright spikes but the dim outer edges will be hard to distinguish from 'real' data and may remain visible. I had this problem with gnarly reflection artifacts introduced by my hyperstar.

I think you should consider those spikes a fact of life for scopes with obstructions. It's merely an aspect of the scope you are considering and in your case it goes into the minus column when comparing different scopes.

Some people pay good money for Noel Carboni's spike software. Same category as spending money on 'fish lips' IMO

[dnault42]

Thanks for the feedback Sander.  Since technically the diffraction spikes are part of the stars PSF it seemed like there might be some energy from the surrounding signal that could be teased out, but if in practice that signal is very small compared to the energy from the spike I can see how that would be trying to get something out of nothing.  Thanks for clarifying.  I'm assuming this is one of the reasons you chose the EdgeHD 1100?

Regards,
David

[Nocturnal]

Hi David,

there were several reasons for choosing the HD 1100 but spikes didn't really come into consideration. I like hyperstar imaging so being able to use the hyperstar from my C11 for my HD 1100 was a plus. The corrector has drawbacks too such as reflection artifacts and reduced cooling of the primary.

I think spikes that come from the optical system are not disturbing and are simply part of the image. I would never add them on purpose to an image that didn't have them because it was shot with a refractor or other optical system that introduces them.

[dnault42]

Sander,

I agree with not artificially adding diffraction spikes.  Seems like wasting data to me, but to each their own.  Thanks for the info.

Regards,
David

[pfile]

- take two or more sets of data with the camera rotated and do a sigma combine to reduce the spikes before post processing. This works for bright spikes but the dim outer edges will be hard to distinguish from 'real' data and may remain visible. I had this problem with gnarly reflection artifacts introduced by my hyperstar.

isn't it the case that the spike angles are a function of how the telescope is oriented toward the sky, rather than the position angle of the camera? when you register the camera-rotated images, they just get rotated right back such that the spike angles are the same.

i'm pretty sure i tried doing this and it did not work. i'm thinking if it were a matter of simply rotating the camera and sigma-clipping away the spikes that everyone with vane-supported secondary mirrors would just do this and spikes would fall into the 'annoyance' category rather than the 'minus' category for this class of telescopes.

[Carlos Milovic]

Yes, it is a function of the orientation of the telescope.

You may have an alt-az mount, with a field derotator, and then the spikes should move.

[Nocturnal]

Doh! Good point! Thanks for the correction.