Egg shaped stars at the edge of the field [Solar System] Processing techniques · mousta · ... · 10 · 519 · 3

mousta 2.71
...
Hi All,
I was hoping that there is a PixInsight way to make egg shaped stars at the corners of the picture round. I tried many things and searched online but no luck
Im using a refractor - I don’t have a flattner for now.  I understand that ideally my images shouldn’t be even producing them if I Have a field flattener and my focus/tracking is good. Is there a way to fix it without the flattner or it’s 100% necessary to buy one ? 

https://www.astrobin.com/b4sujg/

ps (any comments/suggestions on the picture are welcome
Like
Sean1980 3.15
...
·  2 likes
As far as I know no way to fix this in processing. Flattener is a must to get round stars in the entire field of view
Like
barnold84
...
·  4 likes
I also can only encourage you not to look for the "software pill for the optically ill" but to buy a flattener. It's not just that the stars are out of shape they are also out of focus. 

CS!
Björn
Like
jerryyyyy 9.03
...
If you have PixInsight you can try MorphologicalTransformation... in theory... if you cannot fix the optics... you can also crop....
Edited ...
Like
JohnAdastra 1.81
...
Mousta,

That is a tough problem in processing since the stars are more elongated the further into the corners you go. Normally you could increase back focus to change this effect, but without a flattener, you are probably just using the focuser rack to obtain focus, and not doing back focus from any fixed point. 

Deconvolution can be used to fix elongated stars, but since the star shapes are not consistent you can't apply in the whole image. You would need to use a mask to hit each corner with a different Deconvolution orientation, a pretty complex scenario.

John
Edited ...
Like
Krizan 5.73
...
What you are asking is more dificult than you might think.  Which has been noted with above post. I agree with Bjorn, an easy software fix is not you answer. You will have to correct the other issues first. The stars or disorted/elongated differently in each corner. This is because the camera is not orthoganal with the light cone. This can be caused by flexure in the imaging train or the camera chip not being square with the the scopes light cone. Notise that the stars or larger an out of focus in the lower left corner, and the stars are smaller and in focus in the upper right corner. This would require different correction methods in each corner with software.  A flattened will help, but still may not solve all the inconsistent corner stars, until the imaging train is orthoganal with the light cone. 
.

Lynn K.
Like
kajouman 2.81
...
·  1 like
some  very handy info. 915241512_flattenerspacing.jpg.a6833a95acca503166f08bb88911e625.jpg.59fa772a5dc7946306e9f62e8e1a43e0.jpg
Like
barnold84
...
·  3 likes
@John
@mousta is not using a flattener yet. What his problem is is the so-called Petzval field curvature which requires a field flattener. What you are showing is the behavior of incorrect backfocus distance for a flattener/reducer.

@Lynn K I think mousta's major problem and question is the missing correction of fiel curvature. 

@mousta I couldn't help myself and do some mathematical playing. I've elaborated a small transformation that would do a rough fix but that's more for mathematical curiosity instead of a serious problem solving approach, as I'm still with my statement that a flattener is the way to go. If you give me permission, I would show the transformed copy of your image in this thread to show what it would look like?
Like
mousta 2.71
...
Thank you all for the feedback - so much to learn still

@Björn Arnold Thank you so much for your thorough explanation. Of course you can edit/post the image
Like
barnold84
...
·  1 like
@mousta  Thanks. The transformed image at the end of the post. As a general notice: in computer vision, camera systems are calibrated for image distortion. Also in many RAW-Converter programs, there are profiles for DSLR lenses which are supposed to compensate the remaining image distortions. Therefore, I'd guess one could elaborate a software solution. Nevertheless, the goal is always to have corrected optics.

In order to transform the image, I had to find a few parameters by hand (the center of the distortion and the degree of distortion) and therefore, the solution won't be optimal. Judge by yourself. I'm pretty sure that there are software solutions out there which one would find after thorough searches. The problem is probably to place the right keywords in the search box.

Here's the "corrected" image:
transformed.jpg
(the white background should illustrate the original image dimensions)

Here's an overlay of the transformed image with the color negated original image to have another idea how much transformation is necessary:
composed.jpg

There is indeed one thing @Lynn K had mentioned and, contrary to my previous post (my apologies), deserves indeed some attention if you're going to use a flattener: as one can see, the center of the transformation has quite some offset from the image center. I'm not sure if this is a translatorial displacement or sensor misalignment.
Like
wondering_scope 0.00
...
· 
Im using a refractor - I don’t have a flattner for now.


I am probably going to answer this for you some time in 2022. I have tried to mount a full frame Sony a7 mirrorless to a 127ED APO with a 0x7FF/FR and this is a big issue with big sensors. It is why I bought the scope. APC and smaller sensors you can tolerate the edge cases, but with a full frame sensor a flattener is a must have (and I am still working on getting it solved). I have a 0.7x flattener /reducer for the long refractor and will be doing tests up against a plug and play Redcat 71 quadruplet APO ready made for full frame sensors. I hope to get both working but if the 127 fails, the RC71 will be a joy to use anyway.
Like
 
Register or login to create to post a reply.