My Tutorial PixInsight Workflow for DSLR Images

[kayronjm]

Read here: http://lightvortexastronomy.blogspot.com/2013/12/tutorial-post-processing-pixinsight.html

As part of my involvement within the Gibraltar Astronomical Society, we are getting people to experience astrophotography by getting them to use their DSLRs directly mounted on our equatorial mounts, among other things (e.g. telescopes). This made me realise my website is missing a workflow tutorial relating to colour cameras and even more so, DSLRs. This tutorial is an extensive workflow tutorial of the same style I usually do with monochrome CCD camera tutorials, but for DSLRs, with considerations made for the limits of DSLRs in terms of noise. I hope someone, somewhere will find it helpful!

[Zocky]

Excellent tutorial.
It can be applied not just to DSLR users, but to CCD camera users as well.
Well done.

[kayronjm]

Indeed it can, thank you!
The reason I target DSLR users on it because I have other tutorials that are more CCD image centric. But yeah, this tutorial is based on One Shot Colour images, really.

[Batch]

What a great tutorial....thanks for posting it.

Bill

[georg.viehoever]

Excellent! I would suggest to add a table of contents or overview at the beginning, so your readers can get an idea of the overall plan, and directly jump to the steps they are currently interested in.
Thanks a lot,
Georg

[kayronjm]

Excellent! I would suggest to add a table of contents or overview at the beginning, so your readers can get an idea of the overall plan, and directly jump to the steps they are currently interested in.
Thanks a lot,
Georg

This is an excellent idea, thank you. I will do that for all my tutorials as well.

[sctall]

This is an excellent tutorial
I really like the layout and verbosity.
However, I did run into an issue.
I followed it all the way to the LRGB combination.
I saved both images RGB and Lum in 64 bit, but I get this error massage when trying to combine, see attached.

Scott T.

[kayronjm]

This is an excellent tutorial
I really like the layout and verbosity.
However, I did run into an issue.
I followed it all the way to the LRGB combination.
I saved both images RGB and Lum in 64 bit, but I get this error massage when trying to combine, see attached.

Scott T.

Hi Scott, thank you for your comments. The error sounds to me as if the RGB and Luminance images are not the exact same resolution as each other. Did you perhaps do some cropping or resizing to either the RGB or Luminance images AFTER you split them up?

[sctall]

Yes I did, You hit it on the head.

The RGB had a slight dark border I cropped out during the HTT process..
Do you know how to get the images back to the same geometry, without starting over?

I processed this my normal way and was pretty pleased with it, but then I decided to try it your way, and I already see a diff in the RGB image.
I really want to combine the 2 and see what the result is.
I still consider myself a newbie processor with this PI.

Scott T.

[kayronjm]

Yes I did, You hit it on the head.

The RGB had a slight dark border I cropped out during the HTT process..
Do you know how to get the images back to the same geometry, without starting over?

I processed this my normal way and was pretty pleased with it, but then I decided to try it your way, and I already see a diff in the RGB image.
I really want to combine the 2 and see what the result is.
I still consider myself a newbie processor with this PI.

Scott T.

Hi Scott, I'm glad we got to the root of the problem so quickly!
Unfortunately I don't see how you can fix this unless you crop precisely the same amount but then there's no guarantee that everything in RGB and Luminance images matches in terms of overlap. You could try extracting the Luminance from the RGB again but this is unideal because your RGB has already been post-processed. I would personally start over and this time make sure you crop out everything you need to crop out from the original image before extracting the Luminance.

[pfile]

load the history explorer from the RGB image (right click on image, "load history explorer")
open the history explorer; you will see the dynamic crop icon there. drag it out to the desktop.
now drag the dynamic crop process icon from the desktop onto your L image.

should work.

rob

[sctall]

Unfortunately, The image does not have the history.
This could be a downside to saving the image with a new name to keep track of what has been done. Don't know. But it has nothing but the initial state in the History.
I did try to crop the RGB again, and then apply that to the lum, but I had the same issue.
I guess I will be starting over.
Not a bad idea anyway to get more familiar with the process.
I am going to watch the history this time.

Scott T.

[pfile]

well, all is not lost. if you can tolerate registering the L again, simply open StarAlignment, put the name of the RGB image into the reference field, and drag the triangle to the L image. you'll get a re-registered version of the L. hopefully SA will detect that the operation is simply a crop and the image will not be resampled. even if it is, it's fine, but it's best to avoid un-necessary re-registering of images when possible.

image histories are not stored with files written out to disk. but if you instead use Save Project... all the image histories will be saved when you reload the project.

rob

[sctall]

Thanks all

Well I found that if you look at the View Explorer, the stats of your image geometry are all there.
So I looked at both images and with Dynamic crop I typed in the Width and Height for the largest image.
and viola!, I could use the LRGB tool.

So here it is. It could be and I can do better, but tracking wasn't great and with the LP, including the wonderful Christmas lighting, I am surprised it came out this good.

I am open to any suggestions of course.
That is why I like this forum.

On another note:, I actually took a few subs of the target with a shorter exp time to help eliminate some of the saturated core.
But I added all subs into the BPP for this test. I'm sure this isn't correct.
If that is not the best way, could someone explain what would be the preferred method to try to retain the data in the core.


Scott T.

[kayronjm]

Thanks all

Well I found that if you look at the View Explorer, the stats of your image geometry are all there.
So I looked at both images and with Dynamic crop I typed in the Width and Height for the largest image.
and viola!, I could use the LRGB tool.

So here it is. It could be and I can do better, but tracking wasn't great and with the LP, including the wonderful Christmas lighting, I am surprised it came out this good.

I am open to any suggestions of course.
That is why I like this forum.

On another note:, I actually took a few subs of the target with a shorter exp time to help eliminate some of the saturated core.
But I added all subs into the BPP for this test. I'm sure this isn't correct.
If that is not the best way, could someone explain what would be the preferred method to try to retain the data in the core.


Scott T.

Scott, I'm glad you got it sorted out!
I wasn't aware that you'd be able to derive the information from there and therefore apply the same crop - that's excellent. Indeed I am sure that with less light pollution (perhaps with use of a light pollution suppression filter such as the Hutech IDAS LPS), you'd get an overall cleaner image. It does look excellent though!

Regarding your question about retaining information in the core with shorter exposures, you want to create a HDR image. Have a look here:

http://lightvortexastronomy.blogspot.com/2013/02/tutorial-pixinsight-workflow-for-hdr.html#STEP5

That's "step 5" of my HDR tutorial (for the Great Orion Nebula, in particular). Feel free to read the rest to fit it into context but it might suffice for you to just look at "step 5" and what comes after (stretching the histogram but then recovering the detail in the core with other tools).