Multi channels mosaics, how?

[jmtanous]

Hi,

I've been using PI for a couple of years, I am proficient with most of its functions.

In the past I have created mosaics with great success using StarAlignment and GradientMergeMosaic. In my previous mosaics I have mixed the channels before stitching the frames together, so basically I have stitched color images. This works fine for 2 frames mosaics.

Right now I am working on a 3 pane mosaic and I want to stitch the the frames per channel before blending the different channels together. In other words I want to create a 3x1 Ha frame, 3x1 OIII frame and a 3x1 SII frame, then mix the channels using PixelMath.

For instance right now I have a perfectly borderless Ha 3x1 frame. I created the frame using SatarAlignment +  GradientMergeMosaic. How can I create the SII frame?

I was thinking in aligning the individual SII panes to the master 3x1 Ha frame, however I don't know how to match the intensities between the SII frames before stitching them in GradientMergeMosaic.

Is this the best way to stitch a multi channel mosaic?

Any help is welcome!

Cheers,

Jose

[pfile]

I was thinking in aligning the individual SII panes to the master 3x1 Ha frame, however I don't know how to match the intensities between the SII frames before stitching them in GradientMergeMosaic.

this seems like a good strategy to me. to match the intensities, maybe do a linearfit on 2 of the SII frames with the other SII frame as a reference first?

rob

[Tom OD]

I've spent a good bit of time with mosaics in PI with mixed results. I would make a master layer in Luminance, by creating files in Star Alignment and the Mosaic Union setting with frame adaption checked. Using the Star Aligned mosaic, I re input the same Lum original files and get star align to create files with the Mosaic Match setting checked. From these files I put them into Gradent Merge Mosaic to create the new Master Lum mosaic frame. Don't crop it, this is very important. Then use the S2 and O3 files in Star Align and Register Match setting to register those frames to the Master layer, in your case the Ha layer. You should be able to create a Master S2 and O3 layer that's aligned to your Ha layer. I have not tried Linear fit yet, so maybe that helps too before the GMM step.
Tom

[pfile]

ah yes, i had forgotten about the frame adaptation thing, good point.

[jmtanous]

I was thinking in aligning the individual SII panes to the master 3x1 Ha frame, however I don't know how to match the intensities between the SII frames before stitching them in GradientMergeMosaic.

this seems like a good strategy to me. to match the intensities, maybe do a linearfit on 2 of the SII frames with the other SII frame as a reference first?

rob

Rob,

Thanks! Good idea... Not sure if the LinerFit will work with images of different geometry, last time it didn't, for instance the panel1 is slightly different that panel2, the reason I like to crop out the black borders after integrating the images. I can easily overcome this limitation cropping all the images the same amount.

I will give it a try!

Thanks!

Jose

[jmtanous]

I've spent a good bit of time with mosaics in PI with mixed results. I would make a master layer in Luminance, by creating files in Star Alignment and the Mosaic Union setting with frame adaption checked. Using the Star Aligned mosaic, I re input the same Lum original files and get star align to create files with the Mosaic Match setting checked. From these files I put them into Gradent Merge Mosaic to create the new Master Lum mosaic frame. Don't crop it, this is very important. Then use the S2 and O3 files in Star Align and Register Match setting to register those frames to the Master layer, in your case the Ha layer. You should be able to create a Master S2 and O3 layer that's aligned to your Ha layer. I have not tried Linear fit yet, so maybe that helps too before the GMM step.
Tom

Tom,

I am doing something similar, the issue that I am facing right now is how to match the intensities of the SII and OIII panels since the master reference is Ha. I can do it by hand, but most of the time PI has already solved these problems.

Cheers,

Jose

[jmtanous]

this seems like a good strategy to me. to match the intensities, maybe do a linearfit on 2 of the SII frames with the other SII frame as a reference first?

rob

Rob,

I just checked how LinearFit works and it is a pixel by pixel operation, so this tool is designed to be used for registered images or images of the same area of the sky. So It won't work.

Cheers,

Jose

[pfile]

yeah, you are right, i had forgotten about that.

so i guess frame adaptation + gradients merge mosaic is the right way to go.

[jmtanous]

yeah, you are right, i had forgotten about that.

so i guess frame adaptation + gradients merge mosaic is the right way to go.

Is there a way to use/enable frame adaptation for SII and OIII individual frames if my reference image is the Ha linear mosaic?

Cheers,

Jose

[pfile]

i don't see why not, i mean you can select any image as the reference; as long as there are enough good stars in the image SA can register it no matter what the filter was.

[jmtanous]

i don't see why not, i mean you can select any image as the reference; as long as there are enough good stars in the image SA can register it no matter what the filter was.

Rob,

If I use the Ha mosaic as a reference I can register the rest of the panels without any problem. What I don't know is how to match the intensities of the SII and OIII frames. When I stitched together the Ha frames I enabled the 'frame adaptation' option in SA, no problem SA will shift the backgrounds and signal to match between the frames.

When I register the SII against the Ha master there is no way to match the intensities between the SII frames, since the registration is against the Ha mosaic. I can manually match the the background and signal between the SII and OIII frames and then register them against the Ha master but I want to know if there is a better, easier way to do it.

Cheers,

Jose

[pfile]

assuming each filter's subs of each panel were all registered to a common image, i guess you could do the linearfit on each stack against the Ha stack before making the mosaic... right?

also though, if you've made your oiii mosaic panels by registering against the Ha mosaic, in the end don't you have two N-panel images which are registered to each other? can't you then use linearfit between the Ha mosaic and the OIII mosaic in that case?

rob

[Tom OD]

assuming each filter's subs of each panel were all registered to a common image, i guess you could do the linearfit on each stack against the Ha stack before making the mosaic... right?

also though, if you've made your oiii mosaic panels by registering against the Ha mosaic, in the end don't you have two N-panel images which are registered to each other? can't you then use linearfit between the Ha mosaic and the OIII mosaic in that case?

rob

I use SA and GMM for LRGB images, so I do the blending of the frames post registration. I understand your problem of how to make the mosaiced S2 frame adapted also. I don't have any suggestion for your problem. Does registering the S2 frames also try to match the background brightness?

[jmtanous]

assuming each filter's subs of each panel were all registered to a common image, i guess you could do the linearfit on each stack against the Ha stack before making the mosaic... right?

also though, if you've made your oiii mosaic panels by registering against the Ha mosaic, in the end don't you have two N-panel images which are registered to each other? can't you then use linearfit between the Ha mosaic and the OIII mosaic in that case?

rob

Rob,

The problem is the intensity between each panel of the same channel.

When I create the Ha master, the SA match the intensities of each panel every time I stitch them together. For simplicity lets assume that we have a 2 panel mosaic of 2 channels. ha_a1, ha_a2, sii_a1, sii_a2

Now if I stitch ha_a1 and ha_a2 using SA with the 'frame adaptation' option enabled I will get a nice master ha_a1a2 mosaic.

Then if I stitch the sii_a1 against ha_a1a2 I will get an big image sii_a1_registered, half with sii_a1 and the other half blank. The same will happen with the sii_a2, we will endup with one big image sii_a2_registered.

At this stage I can use the GradientMergeMosaic tool to stitch together sii_a1_registered and sii_a2_registere, it will work, however the intensities between sii_a1_registere and sii_a2_registere is not matched and the results won't be optimal.

Assuming the latter workflow, where do you suggest running linerfit?

Cheers,

Jose

[jmtanous]

assuming each filter's subs of each panel were all registered to a common image, i guess you could do the linearfit on each stack against the Ha stack before making the mosaic... right?

also though, if you've made your oiii mosaic panels by registering against the Ha mosaic, in the end don't you have two N-panel images which are registered to each other? can't you then use linearfit between the Ha mosaic and the OIII mosaic in that case?

rob

Does registering the S2 frames also try to match the background brightness?

I don't know. But I guess yes, frame adaptation tries to align the intensities using all the background of the image.

Cheers,

Jose