Cepheus Mega-Mosaic - 40 pane target achieved.

[Adreneline]

Hi Everyone,

For better or worse I have achieved the nonsensical 40 pane target I set myself totalling 41 hours and 9 minutes of Ha.

The object of the exercise was to understand the nebulosity in this region, initially in Ha and hopefully at some stage with OIII and SII, and here it is:

I think I am pushing the limits of the software. Astro Pixel Processor has done a greta job, but not a perfect job; the extent of the curvature has presented a real challenge.

StarXterminator has removed the stars fairly cleanly and with careful processing I have managed to put them back reasonably well. That said the stars shapes are all over the place.

I have tried compiling the mosaic using the individual panes, groups of 2x2 and 3x3 and a mixture of all three.

The result presented below is the best I can get in terms of star shape; there are clearly some distorted stars.

The distortion across this f.o.v. produced all manner of problems when trying to plate solve and it is clear the solving is a bit off in parts of the image - again I suspect due to the curvature.

To put this into context the f.o.v. is approximately 20 degrees by 16 degrees - or ~40 Moon diameters by ~32 Moon diameters.

It is what it is! I am either pushing the limits of the software or I have exceeded my own limits in terms of processing, etc.

I was hoping to get all of NGC 7822 and Sh2-129 but as it is the compiled mosaic is 1.06 GBytes in size and my MacBook Pro has to take a deep breath every time I as it do something with only 16 GBytes of memory.

I hope this is of interest/use to someone who may be imaging this area as part of the current SGL Challenge.

If anyone has any experience of processing very large mosaics like this and knows how to counter curvature, distortion, etc. I am open to all advice and suggestions.

Thank you for looking.

Adrian

P.S. I am going to point the trusty RedCat at something else for a while before I consider going after the OIII and SII.

 

 

[vlaiv]

32 minutes ago, Adreneline said:

If anyone has any experience of processing very large mosaics like this and knows how to counter curvature, distortion, etc. I am open to all advice and suggestions.

Only way is to let software do it for you.

On most panels stars are properly aligned - except in few places.

In order to see if image can be properly aligned - maybe try using Microsoft ICE for making a mosaic. It is much more advanced for that sort of thing as it is made for camera lens and panoramic images. Lens have much more geometric distortion than simple system like telescope (which should not have any distortion except for standard sphere to flat surface mapping issue).

Only problem is that Microsoft ICE works with 16 bit data and that you probably need to feed it stretched images which poses a problem of its own. Maybe best way to do it would be to perform stacking and then identical non linear transform to each stack.

Something like background removal followed by gamma transform (with gamma being in range of 2-3, or even more if it provides better result) and conversion to 16bit.

[pipnina]

Pixinsight's mosaic workflow is rather painful, but pixinsight can help you with distortion correction and i think only needs two images to do it!

Star alignment script, split mosaic mode for example has a distortion correction setting which I have found very effective even when only merging two images.

I don't know if the result afterwards can be used in other programs that will actually be easier to merge mosaics together with... But it could be an option if Pix is part of your workflow already.

Your mosaic looks excellent to me already though!

[Adreneline]

16 minutes ago, vlaiv said:

maybe try using Microsoft ICE for making a mosaic

Spooky. I was out on my daily walk today and thought - I wonder if ICE might sort this. Thanks Vlad. I was determined to use APP as it is such a great piece of software but I think I am pushing the limits of what it can handle. All of my 2x2 and 3x3 sub-panes are stitched perfectly but as soon as I compile the whole things it starts to struggle. If I assemble the bottom half  it works fine but as soon as I join the top half it starts to show signs of 'distress'. I feel it pain!  It's the same if I do left vs. right.

I will experiment and see how I get on. Thanks for the help/advice/suggestions - much appreciated.

Adrian

 

[Adreneline]

7 minutes ago, pipnina said:

Pixinsight's mosaic workflow is rather painful

Compared with APP I am inclined to agree - "painful". Again I did look briefly at PI the other night but then my brain stalled and I thought I need to look when I am more awake!

As mentioned above APP works perfectly on small mosaics but I am pushing the limits here - a 20 degree f.o.v. is a bit extreme. I mentioned 1.06 GBytes above and that is the file size with individual panes down-sized by a factor  of two!. APP is doing an amazing job - no doubt in my mind. That said I will try PI when my brain is fully awake!

Many thanks for your advice and suggestions.

Adrian

 

[geeklee]

Fantastic Adrian ( @Adreneline ).  Love the star filled image the most - a perfect balance of processing the differing parts and bringing them together.

The PI annotated image is just the cherry on top, hammering home just how big an area of sky you've managed to cover in Cepheus.

All the more impressive that it's in UK skies, only the start of October and you didn't start your posts with "Well, it's been 5 years in the making but..."

Look forward to more updates.

Edited October 5, 2022 by geeklee

[Adreneline]

4 minutes ago, geeklee said:

Fantastic Adrian ( @Adreneline ).  Love the star filled image the most - a perfect balance of processing the differing parts and bringing them together.

The PI annotated image is just the cherry on top, hammering home just how big an area of sky you've managed to cover in Cepheus.

All the more impressive that it's in UK skies, only the start of October and you didn't start your posts with "Well, it's been 5 years in the making but..."

Look forward to more updates.

Thank you Lee - and thank you for your feedback and advice throughout the journey. I started this in August and didn't expect to get enough clear skies to complete the Ha this early. I am hopeful I will get the OIII but that said clear, moonless nights are rare in the UK.

Glad you like the star filled version. It is frustrating that the star shape in each individual pane is near perfect but the mosaic has regions where things are clearly amiss. I've tried everything I can think of to overcome the problem  but all it seems to do is move the problem somewhere else! Hence the feeling this has to do with curvature.

"Well, it's been 5 years in the making but..." - what a difference a pier makes!!! Best £500 I ever spent - get one!!! 😉 🤣

Adrian

 

[Fegato]

Amazing job! For some reason I particularly like the lion at the bottom! 

I've only done small mosaics, and do indeed find the standard Pixinsight functionality a bit clunky. However, I do like the Mosaic Scripts - Mosaic by Co-ordinates and Photometric Mosaic. 

[Adreneline]

1 hour ago, Fegato said:

Amazing job! For some reason I particularly like the lion at the bottom! 

I've only done small mosaics, and do indeed find the standard Pixinsight functionality a bit clunky. However, I do like the Mosaic Scripts - Mosaic by Co-ordinates and Photometric Mosaic. 

Thank you Robin, and also for the tips re. PixInsight. I think I need to get the co-ordinates method tried out on my 40 panes; obtaining the centre coordinates for each pane is no big deal as I have them already from the Telescopious planner.

Adrian

[Zummerzet_Leveller]

Great work Adrian, I hope you manage to resolve your mosaic issues.  One of those images that it's easy to get lost in and while away 20mins.

[Adreneline]

3 minutes ago, Zummerzet_Leveller said:

Great work Adrian, I hope you manage to resolve your mosaic issues.  One of those images that it's easy to get lost in and while away 20mins.

Thank you Jody. I am going to have a go with creating the mosaic in PixInsight and see if that can deal with the curvature. Cepheus is such a rich region and not just with the popular targets.

Adrian

[R26 oldtimer]

Fantastic!

[Adreneline]

4 minutes ago, R26 oldtimer said:

Fantastic!

Thank you - I'm glad you like it.

Adrian