M31 Andromeda Galaxy (and
July - September 2013
This is a Narrowband for Stars (nb4stars) project. For more
on the theory of what I am trying please consult the nb4stars page.
Andomeda Galaxy is the most distant object that can be seen with the
naked eye. It rises as the summer Milky Way is setting. Both
Andromeda and the Milky Way belong to the Local Group which is the collection of relatively close galaxies. The
Milky Way, Andromeda Galaxy (Messier 31 or
NGC 224), and M-33 (which
is associated with M31) dominate the Local Group. The rest of the members are small galaxies or even smaller dwarfs.
M31's has a number of satellite galaxies, These are spread over the fall sky. As mentioned
above the 3rd largest galaxy in the Local Group, M33, appears to have interacted with M31 in the past. It is located
in Triangulum some distance from M31. Two other smaller
galaxies (NGC 147 and 185) are located between
Andromeda and Cassiopeia. Two others orbit close to
M31 (M32 and M110). A telescope with a wide field of view
will easily show M32 as a fuzzy ball next to the bright center of M31.
M110 is more of a challenge since its surface brightness is
lower, but can be observed in a dark sky.
I started collecting data for this project right before sunrise in
July. By September I was able to collect for most of the
night. This project used my nb4stars filter set; Strömgren V
and Y filters (which I designate as sV and sYel), a 20 nm Hydrogen
filter, and a 3nm Hydrogen filter. The sV and sYel data were
only collected with the moon fully down. The Hydrogen was
collected with partial moon.
Half Size Image
This image is displayed at half resolution (~ 7"/pixel).
Click on the image above for a
full size image.
The full size image shows the structure of the spiral in more detail.
Note also NGC 206. For comparison here
is a closeup using a larger scope.
image is also presented below.
The first thing I noticed when processing the image was how well the
spirals show. Visually it is tough to see the inner spirals.
Even with a 30" telescope
you can only see the outer spiral and even then it requires a good
night. The 3 nm Hydrogen filter easily collects the star
forming regions so adding it to the base RGB image makes for a much
more dramatic image.
Composing the Images
I completed the processing of this project immediate following the M8 project.
This processing was simple compared to the struggles I had with that
project. All processing was done using Pixinsight 1.8.
Even with 15 minute exposures my narrowband filters collect a very weak
signal. Thus I purposefully collected data until the moon shut
|sV (Strömgren V)
|sYel (Strömgren Y)
|20nm Hydrogen (Ha20)
All filters were Astrodon
Combining the H data
The first step was to form a composite of the Ha20 and 3 nm Ha that
will enhance the Hydrogen rich areas without blowing out the red
portion of the stars. The best method I have found is the Vicent Peris
technique described in this Harry's Astro Shed tutorial.
With my filters the final Red Image was formed in two steps using
An initial RGBpH image is then created by using ChannelCombination to
combine R_VP, sYel, and sV to an RGB image.
- H_VP = ((Hydrogen*20)-(Ha20*3))/(20-3). You can
see the result of this operation here.
- R_VP = Ha20+(H_VP-Med(H_VP))*4
As an example here is a before and after
This shows the rather flat appearance of the RGB before adding Ha.
At this point Color Calibration has not been done.
The Ha20 filter will collect more hydrogen relative to its
bandwidth than a conventional R filter. Thus the star forming
regions are just starting to show.
Adding the 3nm Ha makes the star forming regions pop.
At this point the image is a bit red. That will be corrected
in later steps.
G2V Color Calibration
To insure that stars are the correct colors one images a
reference G2V (i.e. sun like) star. You then balance the
colors so this star appears white. I shot M31 in a wide swath
of sky so no single calibration was going to be perfect. I
tried letting Pixinsight calibrate via structure detection, but that
did not give good results. Thus I used a color calibration
for a star that is high in the sky (since almost all images were shot
about 50 degrees elevation).
For this project
HIP 96037. From this I
determined that the correction should be
These were used for color correction in PixInsight.
The initial color correction worked well with the galaxy mostly a white
color with a slight red tinge. However, after stretching the galaxy got
a yellow tint (too little blue). I removed this using a masked
CurvesTransformation that increased the blue.
How accurate are the star colors? For comparison here is
an version of the image with the BV Index of the stars
annotated. The reader can decide for himself if the colors
Once color calibration was done I used TGVDenoise to reduce
the noise in the background. I was happy with the stars so I
did not bother using deconvolution .
So far the image had been linear. I used a simple HT to
stretch it. I was not happy with the results when I tried a
Bringing Out the Structure
A lot of structure was apparent even at this point.
To make sure the final image did not look over processed I
used the sharping tools very judiciously.The first is
HDRMultiScaleTranform which spreads the brightness over a wider range
of values. I then added a bit of DarkStructureEnhance and
Finally I used MMT to further sharpen the structures the
previous steps disclosed. Then another pass of TGVDenoise to
remove the last noise.
Copyrights for Photos