Digging out the ISM surrounding M45 (better than Photoshop)?

[Josh Lake]

Last week, I took about 4 hours of data on M45. I went too long on the subexposures, so the main stars are blown out, but otherwise I was happy with my results: http://i.imgur.com/oeNgR.jpg

I was pleased that I was able to dig out any of the interstellar medium in the background.

I posted the image and data to Reddit and a fellow user took up the challenge and tried the data. They said they removed the gradients in Maxim DL and then did the rest of the processing in Photoshop CS2. They got a great result: http://i.imgur.com/KrARR.jpg

I have asked for more details so that I can try to replicate the process in PixInsight (hopefully even better), but I wondered if anyone here would be able to dig out the faint white wispies. If so, I'd love to see the workflow that led up to an image similar to the PS user's results.

Here is the data: https://www.dropbox.com/s/c4jhhfxaasewa7d/PIFTW-M45LRGB.zip

[Josh Lake]

I got a better version from the PI processor as well as some detail about what they used: http://www.aeonraven.com/astro/PIFTW-M45LRGB-process2.jpg

The steps I took are as follows:
Open LRG&B files in MaxIm DL. Remove gradients. Save new fits.
Open L in Ps. I like to work on just the L before dealing with any color. So, I import the L fit into Ps and adjust the histogram with curves and levels for an overall good appearance of bright and faint nebulosity.
To increase the visibility of the extremely faint IFN I used an Unsharp Mask with a radius of 250 pixels and a strength of 400% on just the background. This is the step that really did it.
After that it was a bit more independant curves/levels on the brighter and fainter parts to get a good overall balance.
Then I combined the RGB data, solved a few more gradient issues and got the color balanced. Once I was happy with the image as an RGB, I replaced the rgb's L with the processed Luminance file.
From there, it was just getting the color right, making a few more levels adjustments, and reduced it 50% for the .jpg

[pfile]

does this image count?



i'd have to dig up the project, if i saved a project for this one. it's not as deep as your image but some of the IFN stuff is there.

[Josh Lake]

Sure it counts! So how'd you do it? How did you get at the faint stuff?

[pfile]

i'd like to say "i used multiscale processing and pixel math!" but i don't think i did. the truth is that it's 10.5 hours of integration time at f/4 (from a light polluted site). mainly i think it was careful background extraction, which would not have been possible were it not for ferran's insanely deep image here:



basically i registered his image to mine, and then used his image to place the points for DBE and applied that to my image. i'll try to find the project on my server later tonight and then maybe i'll remember more of what i did.

[Josh Lake]

Re: Using someone else's image as a point-placer for DBE -- great idea! I've never tried that.

[pfile]

i guess it's cheating a little bit but it was almost impossible to tell what was signal and what was background in the integrated result. anyway StarAlignment does such a great job that it's easy to register random images to one another. i think i did use previews though to help it out a tad.

[marekc]

This is actually a really interesting idea, using another image to guide the placement of DBE points. If it's described openly in one's account of the processing workflow, with credit given to the creator of the other image, it seems like an `honest' method to me. It's certainly not `painting'!

It would be interesting to look at surveys of the IFN, made by research astronomers... they might provide a `master database for placing DBE points, maybe? Might make a fun project!

- Marek

[pfile]

i think RBA described something like this in one of his widefield talks. i need to go find it now. basically he was consulting some sky survey data to see if his DBE made sense. but as i recall that was kind of low-resolution data, so i'm not sure it could be registered properly.

[Raymond]

Hi
RBA has a description on his homepage:
http://blog.deepskycolors.com/archivo/2010/05/01/luminance-Processing--Making-the-Ifn-p.html
Maybe that helps...

[dayers]

Pfile,

I have been easing my way into plate solving and wonder how  you went about annotating your M45 image? Using a Cygwin shell on Win7, I have been able to solve an image and get the circles around the objects via astrometry.com but not the descriptions. I have a sneaking hunch that, as usual, I am going at it the hard way.

If you can give me a nudge in the right direction I would appreciate it.

Dave

[pfile]

Hi Dave -

i can't remember all the options for the astrometry.net annotator. the M45 that's on astrobin was automatically annotated by astrobin, which also uses astrometry. so you should be able to get the same annotations as salvatore is generating. unfortunately i can't check as my MacPorts installation is all messed up now and plotxy and plotstuff refuse to run.

anyway, lately i've been solving my images with astrometry and then using the annotate script in PI. the only gotcha is, and i have not gotten to the bottom of this, that i have to flip the image across the horizontal axis, save that, and then solve the flipped one. otherwise the annotation is upside-down. this might have something to do with the PI fits writer settings and how astrometry reads fits files. anyway, because of this i have to open the solved image and use the FITS Header tool to copy the keywords into the original image, then run the annotation script.

works pretty well - here's an example:

http://www.flickr.com/photos/pfile/8208421879/

[dayers]

Thank you. I have an Astrobin account and will look into it.

Dave

[Josh Lake]

the only gotcha is, and i have not gotten to the bottom of this, that i have to flip the image across the horizontal axis, save that, and then solve the flipped one.

I love this script and have used it often, but I find that I need to flip my images *vertically*. Weird!

[Harry page]

Hi
Dave

Why do you not use Pixinsight for this

http://pixinsight.com/forum/index.php?topic=3911.0

Harry