Is it possible to combine data from two different set ups?

[5haan_A]

Hi,

I'd like to know if its possible to combine data from two different scopes on different mounts into the same image. 

Ideally I would like to supplement images taken with osc dslr through a redcat with Ha data taken with a completely different scope, mount and camera.

 

Im trying to get my head round what the work  flow for it would look like. I think I can get framing where I would want it to be, but at what point would I add in the stacked Ha data? Do I have to resize images, if so how? Because the fov is different the size of everything is different stars from one setup would be bigger than the other set up. 

 

My biggest concern is combining the two images on top of each other correctly. Is it something I do at the DSS phase or is it something I do in photoshop? My worry with doing it just in photoshop is aligning the two images precisely it seems like it would fiddly to do it by eye. I saw that there might be a potential Pixinsight solution although I'm not sure, I try to avoid PI if possible because it seems so much harder than photoshop.

 

Thanks,

[Eddie Jones]

Hi Shaan_A

If you look at the bottom of DSS you can set up different nights of data into 'groups' - not tried it myself with different scopes - but think it will work!

Edited July 2, 2020 by Eddie Jones
changed image file format

[Adreneline]

30 minutes ago, 5haan_A said:

I saw that there might be a potential Pixinsight solution although I'm not sure

PI does it easily with the StarAlignment process applied to you master frames. If that doesn’t work then DynamicAlignment will do the trick but with more input from you to help it identify corresponding stars in the two images.

Personally I prefer to use AstroPixelProcessor - it makes combining subs from different sessions and/or different setups a breeze. It’s also amazing for mosaics. It’s like DSS on steroids!

Adrian

[johnoelliott]

I've just downloaded a trial copy of AstroPixelProcessor so I can compare it to PixInsight to see which one I  like more.

[carastro]

I do it all the time.  The tricky thing is getting the scopes/cameras looking at the same thing and orientating them the same otherwise you end up losing a whole load round the edge when you have to crop the bits that don't match up. 

As said above there are software that will do it, I just stack mine and re-align and register my images with Registar before combining.

Carole  

Edited July 3, 2020 by carastro

[michael.h.f.wilkinson]

Astro Pixel Processor (APP) does that very easily. I have combined shots of M33 with the EOS 550D and ASI183MC  before, no problem. In APP I either use the multi-session option and stack the lot, or (much easier I find) I let APP compute a weight map for each stack from the individual sessions/instruments, and then stack the stacked linear results with the weight maps loaded. Works fine.

 

[ollypenrice]

APP and PI can do the resizing and re-aligning, though I do it in Registar. However, they do it 'mechanically' so they won't do a good job if there is a big difference in resolution. There's a technique called 'composite imaging' in which areas of interest are imaged at significantly higher resolution and blended into a lower resolution widefield. In this technique a purely mechanical combining of data will produce glaringly obvious differences in star size, star count and residual noise patterns. When I make a composite image I produce a resized, re-aligned high-res image in Registar but only combine it with the widefield in Ps where I can blend it in so as to make the result seamless. The Pixinsite team dismiss this as 'painting,' a taunt which I happily ignore! Here's an example of 0.9 arcseconds per pixel combined with a 3.5 arcsecs per pixel widefield.

https://www.astrobin.com/full/321869/0/

If you already have PI the cheapest way for you would be to fathom out how to do it. That may not be easy, though! As the others have said, APP would probably be easier. Registar is not cheap but I regard it as close to perfect and love it.

Olly

Edited July 3, 2020 by ollypenrice
Typo

[Laurin Dave]

PI can do it, although not always for me, I’ve found that using Thin Plate Splines, Distortion Correction, increasing RANSAC to 8  and using scale differences (ie as you would for a mosaic) helps, also if the image you are registering is a lot smaller than the master create an appropriate sized and positioned Preview and register to that..  When it fails APP comes to the rescue..  here just do the Star Analysis, Register the frames  save them then move to Photoshop, you may need to adjust the Star Count, the Scale Stops and have Same Camera unticked and distortion correction ticked ..  Again as per doing a mosaic.

Nice M42 @ollypenrice..  by coincidence I was playing around with this yesterday on the Veil data we captured last summer and the Eastern and Western Veils I did  when I got back..  I failed miserably in Pixinsight so did it in APP (which took a while whereas PI if it works takes seconds) ..  

Dave

Edited July 3, 2020 by Laurin Dave

[5haan_A]

Some really good suggestions in here. I think I will give APP a go and also look into Registar. Both look like good options to me. Pixinsight is also something that I should learn, but the learning curve looks a lot steeper than PS to me, also I have finally kind of got to grips with PS now. 

On 03/07/2020 at 09:53, ollypenrice said:

I produce a resized, re-aligned high-res image in Registar but only combine it with the widefield in Ps where I can blend it in so as to make the result seamless.

@ollypenrice i think the above is exactly what I want to do. Can you explain a bit more on how you do this, in particular the part in PS when you combine the widefield and high res point of interest. Also why are there taunts doing this technique? Is it just friendly banter or is it something more. For me I just want to create the best possible images with the data that I collect; so if it helps I think why not?
 

[ollypenrice]

6 hours ago, 5haan_A said:

@ollypenrice i think the above is exactly what I want to do. Can you explain a bit more on how you do this, in particular the part in PS when you combine the widefield and high res point of interest. Also why are there taunts doing this technique? Is it just friendly banter or is it something more. For me I just want to create the best possible images with the data that I collect; so if it helps I think why not?
 

If I thought the technique were unethical I wouldn't use it. Like you, I want to produce good images. There is a PI mentality of being 'true to the data' which I find slightly comical since, if you want to be true to the data, don't process it! I know that's a simplification but it's a simplification which contains some true pointers. Anyway I do what I do and those who don't like it don't have to look at it.

I can't cover all the things I do in blending high res with widefield in a post here, but anyone who familiarises themselves with Ps tools can work out how to do it. The key things are these...

- Get the background of the high res to match that of the widefield perfectly. Registar will do its best but use Dodge and Burn tools in Ps to get it spot on.

- Get the colour balance to match.

- Fade in the high res using a feathered eraser so its presence increases towards the area of interest.

- Don't obsess over the need to to have the high res at 100% even in the middle of the area of interest. This is an exercise in compromise.

Olly