The Perils of Widefield Mosaics

[tomato]

Over the run of recent clear nights (but with Moon becoming increasingly intrusive) I attempted a 6 panel mosaic with the SY135/IMX571c to produce an image with both M31 and M33 in the same FOV. I had issues with flats correction which required a lot of artefact removal in the processing stage but the main issue is that  it appears I haven't achieved a uniform level of star capture across the 6 panels. The star only image clearly shows a lower star density in the top right and bottom left panels. Each background panel was 30 x 2 mins exposures with the galaxy containing panels having double the integration time. Of course, the panels were shot at varying elevations and moon illumination and I think that's the root of the problem.  I don't think it is background variation as the starless image (after much manipulation) is reasonably uniform. I can't see any way to fix this other than retake them or reduce the stars in the other panels, which I suppose would balance them up but is a bit counter productive...

 

[ollypenrice]

That's a very interesting problem to have, an observation which might, quite reasonably, encourage you to throw a brick in my direction!   But I say it because, having made huge numbers of mosaics with up to 35 panels, I've never had it. I've had plenty of others, but not that one, yet I see straight away what you're talking about. Just for background info, I'd have a look at Rogelio Bernal Andreo's image of a similar field just to see if you spot anything comparable.

Did you combine the linear panels then stretch? It's nice to do so and sometimes it works, but what to do when it doesn't? This is what I would do: partially stretch all the individual panels till the background sky value is the same in all of them and the histo peak similarly located. While we might want to end up with a background brightness of 23 in Ps units, we might settle for 15 at this stage. Next, look at the star count in the dark panels and their neighbours., especially where they overlap. What happens if you pin the background in the dark panels and stretch only above that? Can you get the star count to resemble that of the neighbours in the overlap regions? If not, you'll have to re-shoot them, I think.

Personally I like attacking these problems in Ps because I can put one panel next to its neighbour and say, 'That one needs to mesh with this one, so what do I need to do?'

Olly

[tomato]

Thanks Olly,

I combined them in their linear state using APP, but I will have a go at individually stretching each panel then combining to try and achieve a more balanced result.
This is the first time I have encountered this problem, last Spring I did a few galaxy mosaics over a number of nights in the Virgo region but had no such problems, but then the overall FOV was much smaller and they were captured with an F7 refractor.

[tomato]

I actually think the problem is that panels 2 and 5 are better quality than the rest, and so when I did star reduction work on the whole mosaic the smaller, tighter stars in these panels were minimised some way to invisibility.

So I followed Olly's suggestion and applied a partial stretch to these panels before combining, not perfect but better, I think, than my first effort. I have posted the final  version in the  Deep Sky section.

 

[Tom OD]

On 13/08/2022 at 19:07, ollypenrice said:

That's a very interesting problem to have, an observation which might, quite reasonably, encourage you to throw a brick in my direction!   But I say it because, having made huge numbers of mosaics with up to 35 panels, I've never had it. I've had plenty of others, but not that one, yet I see straight away what you're talking about. Just for background info, I'd have a look at Rogelio Bernal Andreo's image of a similar field just to see if you spot anything comparable.

Did you combine the linear panels then stretch? It's nice to do so and sometimes it works, but what to do when it doesn't? This is what I would do: partially stretch all the individual panels till the background sky value is the same in all of them and the histo peak similarly located. While we might want to end up with a background brightness of 23 in Ps units, we might settle for 15 at this stage. Next, look at the star count in the dark panels and their neighbours., especially where they overlap. What happens if you pin the background in the dark panels and stretch only above that? Can you get the star count to resemble that of the neighbours in the overlap regions? If not, you'll have to re-shoot them, I think.

Personally I like attacking these problems in Ps because I can put one panel next to its neighbour and say, 'That one needs to mesh with this one, so what do I need to do?'

Olly

I ve had this same issue and it’s ended in a re shoot in a few occasions. It’s worth going through each of your 30 subs and checking the background values. Maybe sone passing thin cloud in a few frames has messed up the overall stack. Certainly shooting with a moon about won’t help, but it might be a rogue frame or two. 
I can’t remember if I found an issue when I had this problem. I can think of three times it occurred, so as I said I m afraid I reshot the panel in question. 
Tom