TheSkySearchers

Thats nice @Sky Ghostly and ethereal with lots of the faint blue showing.

I had a tough time getting the mosaic put together in PixInsight. I tried stitching the 4 panels together, but, the fourth panel never matched the first three. I finally made a two-panel mosaic out of panels 1 and 2. Then did the same for panels 3 and 4. Then, I combined the two two-panel mosaics into one four panel mosaic. Not sure what the problem was. After that the processing was pretty straight forward. Thanks Bob for the interesting data!
Steve

STEVE333 wrote: ↑Sat Dec 04, 2021 11:28 pm I had a tough time getting the mosaic put together in PixInsight. I tried stitching the 4 panels together, but, the fourth panel never matched the first three. I finally made a two-panel mosaic out of panels 1 and 2. Then did the same for panels 3 and 4. Then, I combined the two two-panel mosaics into one four panel mosaic. Not sure what the problem was. After that the processing was pretty straight forward. Thanks Bob for the interesting data!



Steve

Steve, I had the same issue with ASTAP's mosaic tool, it wouldn't do all 4 panels at once. I wound up with the same sequence as you, 1 and 2, then 3 and 4, then putting the two results together. I think part of the issue had to do with the odd star shapes I got outside the center of each image. I found that cropping away some of the overlapping area in each panel was useful. I shot with 20% overlap and cropped about 7% from each side.

Interestingly, Microsoft ICE had no issue at all with 4 panels in one pass, and it was fast, putting the whole thing together in less than 15 seconds!

In any case, you got a nice result, thanks for taking a shot.

Bob

Sky wrote: ↑Fri Dec 03, 2021 3:30 am Here's my rendition ... completed 100% using Photoshop, except for the use of Microsoft ICE for the mosaic.

You did a really great job of getting all the nebulosity out of the data, Sky! Pulling the stars way back helped with that, but I'd say your process steps were much more important than just reducing the stars was . Thanks for having a go at it.

Bob

bobharmony wrote: ↑Sun Dec 05, 2021 4:10 am
Steve, I had the same issue with ASTAP's mosaic tool, it wouldn't do all 4 panels at once. I wound up with the same sequence as you, 1 and 2, then 3 and 4, then putting the two results together. I think part of the issue had to do with the odd star shapes I got outside the center of each image. I found that cropping away some of the overlapping area in each panel was useful. I shot with 20% overlap and cropped about 7% from each side.

Interestingly, Microsoft ICE had no issue at all with 4 panels in one pass, and it was fast, putting the whole thing together in less than 15 seconds!

In any case, you got a nice result, thanks for taking a shot.

Bob

Thanks Bob, and thanks for the feedback on your problems with combining the four panels. I feel better just knowing it wasn't only me! - Chuckles (Misery likes company or something like that!)

STEVE333 wrote: ↑Sat Dec 04, 2021 11:28 pm
After that the processing was pretty straight forward.



Steve

For a straight forward processing job .... you really nailed it!
Beautiful ... I really like it!

Sky wrote: ↑Sun Dec 05, 2021 3:21 pm

STEVE333 wrote: ↑Sat Dec 04, 2021 11:28 pm
After that the processing was pretty straight forward.



Steve

For a straight forward processing job .... you really nailed it!
Beautiful ... I really like it!

Thanks

The "straight forward" processing was actually more than I had remembered. In PixInsight it consisted of:

LINEAR IMAGE

• MultiscaleLinearTransform noise reduction

• BackgroundNeutralization

• ColorCalibration

• HistogramTransformation (HT)- stretch to non-linear

• TGVDenoise

• HT - adjust black level and stretch a bit more

• Starnet - create starless image along with star_mask (renamed to Stars)

STARLESS IMAGE

• CurvesTransformation (CT)- S-Curve to increase contrast

• CT - Boost color saturation

• LRGBCombination - Extracted Luminance (L) image from Starless image. Performed Surface Blur Noise Reduction (I discussed this technique in the Image Processing sub Forum) on L to reduce background noise. Sharpened noise-reduced L. Used sharpened L as luminance for LRGBCombination. Blurred Starless image before applying LRGBCombination to reduce noise in final image.

COMBINE IMAGES

• Used PixelMath to add Stars to LRGB image. Added Stars at only 60% to reduce star field brightness.

• EZ Star Reduction to further reduce star field brightness.

I guess it wasn't as "straight forward" as I remembered!

Steve

That's an excellent go at it Steve.

You have bought out the colours really well. You have two vertical areas of something else down your image, are these from the mosaic merge?

Regards

Graeme

STEVE333 wrote: ↑Sat Dec 04, 2021 11:28 pm I had a tough time getting the mosaic put together in PixInsight. I tried stitching the 4 panels together, but, the fourth panel never matched the first three. I finally made a two-panel mosaic out of panels 1 and 2. Then did the same for panels 3 and 4. Then, I combined the two two-panel mosaics into one four panel mosaic. Not sure what the problem was. After that the processing was pretty straight forward. Thanks Bob for the interesting data!



Steve

i've been working on it for about an hour and same thing, panel 4 , manualy balancing the histogram for it helped a bit .. i'm using gradiant merg mosiac .. it's like pixinsight assumes all panels have the same signal and just has a fit when it tries to balance 4 or more panels at once , i might give it a go later with doing it in pairs then merging 2x 2 panels for a the 4 panels ....

typicaly when this happens in my 4 panels, i re shoot the panel that has the lower signal and it fixes it, but now it's good to know that there is a way to balance it out . cheers

Graeme1858 wrote: ↑Mon Dec 06, 2021 6:49 am That's an excellent go at it Steve.

You have bought out the colours really well. You have two vertical areas of something else down your image, are these from the mosaic merge?

Regards

Graeme

Thanks Graeme - As I mentioned I had a lot of difficulty just assembling the mosaic. I only used ABE to remove any gradients from each individual panel. When I combined them I didn't do any ABE or DBE on the entire mosaic. That may be part of the problem.

I just tried the following to "fix" the problem: I used the Game script to create Multi point masks that outlined the regions that were brighter. The first picture below shows the masks being created with the Game script. The second picture shows the mask after a little deconvolution to soften the edges.
The mask was applied to the image and then CurvesTransformation was used to adjust the Blue and Green channels to "fix" the regions that were too bright (light bluish). The result is shown below. It's certainly not perfect, but, I think it is better. Of course, this is only a band aid approach. Those problem areas should have been corrected while the data was still in the Linear stage. Wish I had seen them earlier.
I hope this is the issue you were asking about.

Steve

yobbo89 wrote: ↑Mon Dec 06, 2021 4:33 pm

STEVE333 wrote: ↑Sat Dec 04, 2021 11:28 pm I had a tough time getting the mosaic put together in PixInsight. I tried stitching the 4 panels together, but, the fourth panel never matched the first three. I finally made a two-panel mosaic out of panels 1 and 2. Then did the same for panels 3 and 4. Then, I combined the two two-panel mosaics into one four panel mosaic. Not sure what the problem was. After that the processing was pretty straight forward. Thanks Bob for the interesting data!



Steve

i've been working on it for about an hour and same thing, panel 4 , manualy balancing the histogram for it helped a bit .. i'm using gradiant merg mosiac .. it's like pixinsight assumes all panels have the same signal and just has a fit when it tries to balance 4 or more panels at once , i might give it a go later with doing it in pairs then merging 2x 2 panels for a the 4 panels ....

typicaly when this happens in my 4 panels, i re shoot the panel that has the lower signal and it fixes it, but now it's good to know that there is a way to balance it out . cheers

Looks like many of us have had problems assembling the mosaic. I was glad to find a work-around. It will be nice if it works for you with your future mosaics.

Look forward to seeing your version of this challenge.

Steve

This is interesting. I could not get it to stitch well at all in APP but following the process in Light Vortex of making a rough mosaic, registering all panels to the rough mosaic then Gradient merge mosaic I had no trouble whatsoever. I did a crop and ABE (cropped only to ensure the ABE worked) of the 4 individual panels then built the mosaic.

This is the tutorial I followed.

https://www.lightvortexastronomy.com/tu ... osaic.html

Jockinireland wrote: ↑Tue Dec 07, 2021 2:35 pm This is interesting. I could not get it to stitch well at all in APP but following the process in Light Vortex of making a rough mosaic, registering all panels to the rough mosaic then Gradient merge mosaic I had no trouble whatsoever. I did a crop and ABE (cropped only to ensure the ABE worked) of the 4 individual panels then built the mosaic.

This is the tutorial I followed.

https://www.lightvortexastronomy.com/tu ... osaic.html

Hi David - Thanks for the reference to the tutorial. I'll go back and give it a try. My problem was I wasn't able to construct the rough mosaic because I couldn't get the 4th panel to add to the first three.

Steve