StarTools

Dear all,
after some months of very busy work, the lockdown gave me the chance to spend a few hours with astrophotography.
Unfortunately I have some city lights very close and they get into the pictures, as you can see in the linked M13.
Is there anybody so kind to help me with this image? I would really love to see the workflow in the log so to learn and apply it in future similar cases.
I know it is a rather poor image, the location is not good at all and the telescope must be collimated, I apologise

Thank you in advance for your time.
Gaetano

Image FITS can be downloaded from here:
https://www.dropbox.com/s/9zaqjcexvlge0 ... .fits?dl=0

Hi,

Is this a stack or just a single frame?
Would you be able to provide some acquisition details (exposure times, flats, darks, bias, etc.). Thank you!

Hi Ivo,
Sorry, I should have provided such information.

It is a stack assembled by Astap and includes light frames as well as dark frames and flat. No other processing has been performed, just calibration and stack.
I don’t have the full details here with me but all the frames were taken by a KAF8300 based camera, Atik 383L+, and the lights are 120s each.

Gaetano

Much appreciated Gaetano!

Some observations;

• I'm seeing some uneven lighting that does not seem to be corrected by your flats and does not seem celestial in origin (it undulates very fast). It could be too, however I'm seeing an over-corrected dust donut as well at ~ 5 o'clock. Any idea what might have causd this?
• I'm not sure if it's an ASTAP thing, but it appears over-exposed areas (e.g. star cores) exhibit values that are outside of the range specified in the FITS file. I think I have a slight bug in StarTools (i should clip the image properly), but I haven't seen this happen before. Will investigate.
• Your stars are affected by coma.
• It doesn't appear you dithered between frames - after flats it is your #1 way of improving your datasets!

Given your stars are affected by coma, the first thing I do is to correct the field with the Lens module. A side effect of this is that it warps diffraction spikes, however we'll be cropping out those stars anyway.
Because you did not dither, the Lens module will have caused some cool anti-aliasing effects in the background due to the pattern noise. Pattern noise affects so many things and algorithms...
The background and gradients clean up reasonably well with the Vignetting preset. Defects remain, but don't forget the courtesy AutoDev in Wipe is configured specifically to make them visible to you - you can/should always stretch the image less to hide any remaining defects, especially if the object of interest is not image-filling (as is the case here).
So, the simple workflow is this;
--- Auto Develop
To see what we got (see above).
--- Lens
Parameter [Curvature Linked] set to [138.28 %]
--- Crop
Parameter [X1] set to [151 pixels]
Parameter [Y1] set to [604 pixels]
Parameter [X2] set to [1647 pixels (-1021)]
Parameter [Y2] set to [1601 pixels (-407)]
Image size is 1496 x 997
--- Wipe
Vignetting preset.
Parameter [Dark Anomaly Filter] set to [2 pixels]
Parameter [Aggressiveness] set to [96 %] (backing off a little)
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [4.4 pixels] to make AutoDev ignore fine shot noise and pattern noise.
Parameter [Outside RoI Influence] set to [15 %]
Parameter [RoI X1] set to [659 pixels]
Parameter [RoI Y1] set to [415 pixels]
Parameter [RoI X2] set to [879 pixels (-617)]
Parameter [RoI Y2] set to [618 pixels (-379)]
Parameter [Detector Gamma] set to [1.00]
Parameter [Shadow Linearity] set to [49 %] Deconvolution
--- Deconvolution
Auto-generated conservative mask
Parameter [Primary PSF] set to [Moffat Beta=4.765 (Trujillo)]
Parameter [Primary Radius] set to [1.9 pixels]
Parameter [Iterations] set to [22]
Parameter [Mask Fuzz] set to [22.5 pixels]
--- Wavelet De-Noise (switch Tracking off - chose Grain Removal)
Parameter [Grain Size] set to [6.0 pixels]
--- Wavelet De-Noise
Parameter [Smoothness] set to [71 %]

Hope this helps!

First of all, thank you!

Coming to your comments, the uneven light cast from ~3 o’clock could be due to the Moon which was almost full. I captured the flats with the scope in a different position, this could lead to this lack of attenuation, does it sound reasonable?

About the big donut, I don’t know, sorry, it is there and I have not yet understood where is the dust grain causing it. I hoped the flats could get rid of it but it didn’t happen, how can it be?

Coma is there, it is a Newton scope and I have no coma corrector

I didn’t dither, I am learning now how to use this equipment, this is going to be next step How did you realize the frames are not dithered?

I will study the steps from the log ad I am sure I will learn a lot. I have another frameset which is a little better on the same subject, I will try with that one to follow the same strategy.

As at the beginning, thank you!

gaetano wrote:Coming to your comments, the uneven light cast from ~3 o’clock could be due to the Moon which was almost full. I captured the flats with the scope in a different position, this could lead to this lack of attenuation, does it sound reasonable?

Aha! It is indeed possible stray light from the moon made it into your scope. The gradient is just very fast undulating (meaning it changes rapidly over a small area) and at that magnification is not something you would reasonably to be a natural phenomenon. Imaging without the moon should prove this theory.

About the big donut, I don’t know, sorry, it is there and I have not yet understood where is the dust grain causing it. I hoped the flats could get rid of it but it didn’t happen, how can it be?

It happens - it's very likely new dust that was present when the flats were taken, but not when the lights were taken.

Coma is there, it is a Newton scope and I have no coma corrector

The Lens module will help somewhat. Always use it on you uncropped stack first (as on of the first things you do), and only crop the image to whatever dimensions you need afterwards.

I didn’t dither, I am learning now how to use this equipment, this is going to be next step How did you realize the frames are not dithered?

Not dithering causes pattern noise. It is usually visible as streaks. If hot pixels are present they will be dotted around in the same pattern.
Any more sophisticated algorithm absolutely hates pattern noise, as it is forms... well... patterns that are not random Most algorithms are built to enhance or preserve patterns in the face of random noise. Therefore they will not work optimally if pattern noise exists.

Clear skies!

Thank you so much for your time!

For my understanding, how is the Lens module actually working? If I am not wrong, coma is not a field distortion, it is an object one, so the geometry of the FOV should not be changed but the action should be performed on the stars only. I imagine it like a variable deconvolution. Is it correct?

Clear skies!
G

gaetano wrote:Thank you so much for your time!

For my understanding, how is the Lens module actually working? If I am not wrong, coma is not a field distortion, it is an object one, so the geometry of the FOV should not be changed but the action should be performed on the stars only. I imagine it like a variable deconvolution. Is it correct?

Clear skies!
G

The Lens module is based on the work of Brown (1966). As such, it only corrects for distortions. The most common/useful distortion correction is Barrel distortion correction (e.g. where magnification decreases with distance from the optical axis).

Coma is a variation in magnification over the entrance pupil. As such it can be mitigated or eliminated by choosing a corrective curvature/distortion. This attempt to corrects the varying magnification and restore the image.

Does that help?