[Solved] 2 panel mosaic: error

@ippiu,

Thank you very much for sharing your issue.

Did you try normal registration mode instead of mosaic, for a 2 panel, the mosaic mode is actually not needed. Does that work?

And I think that I will need to see your data to fix this problem.

Can you upload your data here:

https://upload.astropixelprocessor.com/

username: uploadData

password: uploadTestData

Please make a folder with your name and issue like : Ippiu-2panelMosaic-error

And upload your data there and let me know once done 😉

I have already opened an issue on my todo list to fix this.

Thanks,

Mabula

@mabula-admin 

Yes, it works with normal registration mode.

But why did you change mosaic mode registration? 

So can you explain better please what you mean with mosaic mode? And when do i need to use it?

Thanks

Andrea

Hi @ippiu,

Nothing has changed with mosaic mode in this regard. Mosaic mode is only needed when 1 or more panels have zero overlap with the main reference frame. With a 2 panel, there will always be overlap with the reference. So it feels like a mosaic, but can still be processed in normal mode.

The error that you get when using mosaic mode still needs to be solved, since it should definitely work with mosaic mode of course. Can you upload the 2 panels if you have not done so already, so I can make sure it works properly in the next release ?

Thanks,

Mabula

@mabula-admin 

Here you can find 2 Ha panels to test yourself: https://upload.astropixelprocessor.com/sharing/kCAtcqc5o

Thanks so much

Andrea

Hi - I am getting the same error with version 45 when performing a two panel join using mosaic mode.  It was working in version 41 which was the last version I used.  Yes it works in Normal registration mode but it also used to work with Mosaic mode.

Posted by: @ippiu

↑

@mabula-admin 

Here you can find 2 Ha panels to test yourself: https://upload.astropixelprocessor.com/sharing/kCAtcqc5o

Thanks so much

Andrea

Dear Ippiu @ippiu,

Thank you very much, I will test and fix it in the coming days, will report back once done.

Mabula

Posted by: @scott44

↑

Hi - I am getting the same error with version 45 when performing a two panel join using mosaic mode.  It was working in version 41 which was the last version I used.  Yes it works in Normal registration mode but it also used to work with Mosaic mode.

Thank you very much Callum @scott44 for reporting this, it sounds then like a more general bug that crept in there... I will test and fix it as soon as possible.

Mabula

Hi Ippiu, @ippiu

Can you double check the frames that you uploaded? 1 frame is not correct I think?

And also for Callum @scott44

I have found and fixed the issue in the meantime using other 2-panel data.

2.0.0-beta will have this fixed, a 2-panel mosaic will work in mosaic mode 😉

Thank you very much for reporting the issue. I will try to release beta46 in the coming days.

https://www.astropixelprocessor.com/community/release-information/astro-pixel-processor-2-0-0-beta46-release-notes-work-in-progress/#post-34752

• FIXED 2-panel mosaic works now in mosiac mode

  As reported here https://www.astropixelprocessor.com/community/main-forum/2-panel-mosaic-error/#post-34674 it you would try to register a 2-panel mosaic in the mosaic mode, it would error with an ArrayIndexOutOfBoundsException. The data could still be registered though in normal mode. The mosaic mode works now as well.

Mabula

@mabula-admin 

Thanks mabula

@mabula-admin thanks kindly Mabula 👍

Hi @ippiu & @scott44,

Please check 2.0.0-beta46, which can again perform the 2 panel mosaic registration mode 😉

Mabula

@mabula-admin 

thanks

@mabula-admin if I understand,  i can do my 12 panels also in normal registration? They all got 10% overlapping.

Stefan (@stefanastro)

No - that's not going to work with 12 similar sized panels. They would all have to overlap with the reference frame. You could possibly do it with 9 panels assuming the centre panel overlapped the other 8, but I doubt I would attempt that.

JC

@connor231 hmm.. maybe I can do 4 of 3 panels. Load each time 3 panels of each row and intergrate that. After that load the 4 rows in there and try again or in mosaic mode.

Because if I load the 12 Ha and 12 OIII panels and do the mosaic mode, it takes 4hours?

I got a pc with 32 core and 128gb ddr3 speed, but that it takes so much time.

It need 278 2-views?! 🙄.

Should I better stack HA-OIII with each panel.  I use HA-OIII dualband filter with asi2600MC-P. 

If I got a color HA-OIII panels and do the mosaic mode and after intergation, using Multi-NB extract HA +OIII

Stefan (@stefanastro)

I definitely would try and combine the Ha and Oiii before trying the mosaic registration. Ideally you would try to register only 12 files, which should be a lot faster.

I have been coming more to the view that normalisation, rgb combine and probably LPR should be done before mosaic registration, if only so that MBB can work well with less seam issues.

But there is probably no single answer that works well in every case.

JC

@connor231 

Hi, i am going to stack each panel on multi-NB color from the color settings (Tab 0).

The pc is busy now to stack the panels now in color, so after that I'll do the LPC and than reload the panels for mosaic and when that is done I load the giant image and do extraction.

I'll let you know how the results are.

I already have the mosaic done in HA and OIIl and combine it. But i  the futher this mosaic image will be come bigger, it is a 36 panel mosaic.

I want the hole Cygnus,  but already got 13 panels.

Not all the panels get 10h, panels with less nebula gets 6h and the area with a lot of gasses 10h or where OIII is 12-15h.

I'll show it later

Posted by: @connor231

↑

Stefan (@stefanastro)

I definitely would try and combine the Ha and Oiii before trying the mosaic registration. Ideally you would try to register only 12 files, which should be a lot faster.

I have been coming more to the view that normalisation, rgb combine and probably LPR should be done before mosaic registration, if only so that MBB can work well with less seam issues.

But there is probably no single answer that works well in every case.

JC

This is my image of 2025, 9 panels. Each got 3a5 hours

The new one got 12 and I gave all panels 10 hours. The sadr region, the SNR and where you have the WR-134 more than 13 or even 15h ours for the OIII this year.

Now today I did the test, stacked each panel as Multi-Nb color, used Remove light pollution.

Load the panels and did the mosaic registration, it had 66 2-views to do and tooks 1h. Did the  intergration.

After that load the large image to extract HA and OIII, but no....after intergration it was still color.

So I checked the difference between the extract version that I used in Combine RGB tool with the color version. 

It's better to extract each panel first before you do the mosaic process.

The reason is that you can play more with the colors in Combine RGB tools. The OIII shows up more out of the image than the color mosaic process that I tested today.

So even the mosaic proces can takes some hours, it gives better results.

So i save my 2 large images of HA and OIII filter so I can load in the future more panels in there. I want the hole constellation Cygnus.

Got 23 panels to go, most panels of those that I saw got less nebulas.

Stefan (@stefanastro)

Looks like the mosaic integration has worked very well - no seams at all. I think you are well on your way to an impressive image - congrats.

JC

@connor231 thank you.

The best way is still extracting.

Have a clear skies