Has anybody masked foreground stars from image of M31?

When we look at the Andromeda Galaxy, we see a mix of light from
the galaxy and vicinity itself, and light from foreground objects
such as stars within our own galaxy. To what degree have these
sources of light been classified as to which galaxy they belong to?
For example, is there somewhere a complete list of all foreground
(Milky Way) stars in front of the main disk of M31 and for some
reasonable distance outward from there? If so, has anyone
undertaken the task of producing a modified M31 image where all
foreground stars have been masked out, thus producing a *true*
image of M31 all by itself?

Robert Maas, http://tinyurl.com/uh3t wrote:
When we look at the Andromeda Galaxy, we see a mix of light from
the galaxy and vicinity itself, and light from foreground objects
such as stars within our own galaxy. To what degree have these
sources of light been classified as to which galaxy they belong to?
For example, is there somewhere a complete list of all foreground
(Milky Way) stars in front of the main disk of M31 and for some
reasonable distance outward from there? If so, has anyone
undertaken the task of producing a modified M31 image where all
foreground stars have been masked out, thus producing a *true*
image of M31 all by itself?

That's like asking for the whole truth and nothing but the truth.
Good luck with that.

~ BG

On Nov 21, 2:29*pm, (Robert Maas,
http://tinyurl.com/uh3t) wrote:
When we look at the Andromeda Galaxy, we see a mix of light from
the galaxy and vicinity itself, and light from foreground objects
such as stars within our own galaxy. To what degree have these
sources of light been classified as to which galaxy they belong to?
For example, is there somewhere a complete list of all foreground
(Milky Way) stars in front of the main disk of M31 and for some
reasonable distance outward from there? If so, has anyone
undertaken the task of producing a modified M31 image where all
foreground stars have been masked out, thus producing a *true*
image of M31 all by itself?

I suppose then you would also like for the *background* to be
removed?

With enough information it would be possible to distinguish foreground
stars and background galaxies by their colors, magnitude, redshift,
proper motion, etc. Whether or not this has actually been done to
make a pretty picture -- rather than just to categorize the objects
for scientific study -- I don't know,

Craig

Craig wrote in
:

On Nov 21, 2:29*pm, (Robert Maas,
http://tinyurl.com/uh3t) wrote:
When we look at the Andromeda Galaxy, we see a mix of light from
the galaxy and vicinity itself, and light from foreground objects
such as stars within our own galaxy. To what degree have these
sources of light been classified as to which galaxy they belong to?
For example, is there somewhere a complete list of all foreground
(Milky Way) stars in front of the main disk of M31 and for some
reasonable distance outward from there? If so, has anyone
undertaken the task of producing a modified M31 image where all
foreground stars have been masked out, thus producing a *true*
image of M31 all by itself?

I suppose then you would also like for the *background* to be
removed?

With enough information it would be possible to distinguish foreground
stars and background galaxies by their colors, magnitude, redshift,
proper motion, etc. Whether or not this has actually been done to
make a pretty picture -- rather than just to categorize the objects
for scientific study -- I don't know,

It's a good question, because you forgot the part where it would be
useful for producing a high contrast image, uncontaminated by "light
pollution" from foreground stars.

The answer to his question is "probably not," since techniques probably
don't exist for masking out more than one star. It hasn't been that
long since interferometry and coronagraphs were state of the art, and
the OP is asking for an even more advanced method. The mathematics of
doing this would probably be enough to keep five PhD's busy for the next
twenty years.

If the OP is serious, he should make a catalog of all the stars blocking
the images of all the easy to find galaxies, to see if there's a
situation where the contrast of one nearby galaxy could be boosted by
eliminating just a single star. This should be possible by adapting the
existing techniques. If so, this idea is definitely worth a proposal
for preliminary study, and the OP could suddenly find himself a PI and
an employed scientist, rather than just a speculative amateur, sitting
at his PC and inventing wild ideas.

The difference between a professional and an amateur is not just having
a good idea. It's having a good idea that can be made into a real
research project that get published. This could probably be done, but
it would take work. Maybe there's a needy PhD student out there looking
for a dissertation project.

With enough information it would be possible to distinguish foreground
stars and background galaxies by their colors, magnitude, redshift,
proper motion, etc.

Yes, that is my idea as information going into this process. I've
been under the impression that Sloan and or Hipparchus contains a
listing of all point sources visible in medium-sized telescopes,
with enough data about each to make such a classification.

Then semi-accurate info about the sky coordinates and scale and
orientation of some old image, plus modern image-processing
software to match features between the image and a sky map, ought
to produce a near-exact mapping between image pixel coordinates and
sky coordinates, whereby the pixel coordinates of each cataloged
object can be calculated. Then an algorithm can verify that the
point source is visible in the image, and remove by averaging it
down to the local context so that it shows as a gray smudge instead
of a bright spot. It might even be possible to gray-out the
diffraction spikes around the spot.

Whether or not this has actually been done to
make a pretty picture -- rather than just to categorize the objects
for scientific study -- I don't know,

I posted to find out if anybody knew if this had actually been
done, at least for some of the common images such as M31 that are
so very commonly posted for public view. (I also posted it as a
challenge to anyone with access to all the data (and understanding
how to use it), and access to suitable image-processing software
(ditto), in case they might say "AFAIK it was never been done
before, but after you posted I did it!!".)

From: Bluuuue Rajah Bluuuuue@Rajah.
It's a good question, because you forgot the part where it would
be useful for producing a high contrast image, uncontaminated by
"light pollution" from foreground stars.

Yeah, that was most of my idea: When we look at an image such as
M31, we see lots of bright spots of light. We can't tell from
looking at the image which are foreground stars having nothing to
do with M31 and which are bright stars or globular clusters within
M31. This makes it very difficult to fully understand what we are
viewing. If we see something that looks out of place, we have no
way to know whether it's a feature of M31 that we need to
understand, or just a foreground star that we ought to not be
seeing in the image except for our location within our own galaxy
with hoardes of stars interfering with our view of M31.

The answer to his question is "probably not," since techniques
probably don't exist for masking out more than one star.

Perhaps you misunderstood my idea. I'm talking about
post-processing of the already-collected (archival) digital image
to remove artifacts including foreground stars, not physically
placing a occluding disk in the focal plane when making a new
image. Surely if you know the locations of ten foreground stars
within the archival image, and you can process the image to
"eliminate" one of the stars, you can repeat the process for each
of the other stars just the same.

If the OP is serious, he should make a catalog of all the stars
blocking the images of all the easy to find galaxies, to see if
there's a situation where the contrast of one nearby galaxy could
be boosted by eliminating just a single star.

That's not my purpose. I expect removing the foreground stars from
an image won't significantly reduce the peak pixel intensity, thus
won't permit ramping up the contrast much more than it was set
already to achieve maximum span from black to white. Maybe some
single foreground star is brighter than the peak central brightness
of the core of M31, maybe not.

Also, I don't know how to get access to the key data I would need
to determine how many known foreground stars are visible in some
region around M31.

This should be possible by adapting the existing techniques. If
so, this idea is definitely worth a proposal for preliminary study,
and the OP could suddenly find himself a PI and an employed
scientist, rather than just a speculative amateur, sitting at his
PC and inventing wild ideas.

I really don't think this idea would warrant a professional
research project. But it might be useful for some company that
manufactures telescopes and sells a package of image processing
software and GoTo control for the telescope. It could brag that the
GoTo database includes all of the catalogs that exist to date, and
that it can automatically find all known objects within any region
of interest, and can edit images according to several algorithms
using that data, such as masking out foreground stars to gray (to
make them go away) or purple (to make them stand out like "sore
thumbs"). The advertisement could then show comparative images of
M31 with and without masking out of foreground stars. It could brag
"this is how M31 normally looks when viewed from Earth, and this is
how M31 *really* looks without our own galaxy blocking the view".

Another possible type of person who might be interested in my idea
are the people who submit their best images to the Gallery section
of Sky and Telescope. I'd like to see somebody post an image of M31
with foreground stars removed. If S&T can post images of landscapes
at night with polar star trails in background, and images of an
eclipse viewed through pretty clouds, and images of sun dogs, I
think there's also room to post a galaxy image with foreground
stars removed. Once such an image gets constructed and published,
then maybe one of the advertisers would adopt the same idea in
their ads.

OT: Another neat thing I'd like to see is a stereoscopic view of
common asterisms such as the coathanger which have been proven as
chance line-ups of unrelated stars of varying distances. Or maybe a
3-stereoscopic view, with a central image just the way the asterism
looks from Earth, and views to each side showing how the asterism
would look as viewed a few lightyears to each side, so you can
choose either two adjacent images to get a stereoscopic view from
that side of the Earth line of sight. Or maybe a hologram that
shows true (scaled) distances which can be viewed from any
direction with any orientation.

The difference between a professional and an amateur is not just
having a good idea. It's having a good idea that can be made
into a real research project that get published. This could
probably be done, but it would take work. Maybe there's a needy
PhD student out there looking for a dissertation project.

Ignoring the fact that I don't think my idea would have any
significant true scientific value, only advertising value or
show-off value or public-education value (helping the public to
understand what a galaxy really looks like, without the distraction
of the foreground stars giving a wrong impression of the galaxy),
there's the fact that I don't have the English writing skill to
write a research or R&D proposal even for the kinds of projects
that I truly feel would be worth pursuing for scientific or
technological or society-improvement value. For example, see my Web
page listing several projects in Web software that I'd like to find
somebody else to work with me to brainstorm the design and help
with testing even if I do all the actual work.
http://shell.rawbw.com/~rem/WAP/projectIdeas.html
I have no idea how to find anyone interested in working with me on
any of those projects, and I have no idea how to write up any of
them as a formal proposal that anyone would pay serious attention
to. And even if I did take a chance at writing a formal proposal, I
have no idea how to find anybody else to proofread it to show me
all the stupid mistakes I made so that I can make the proposal
halfway decent before I submit it to some funding agency.

(Robert Maas, http://tinyurl.com/uh3t)
wrote in :

With enough information it would be possible to distinguish
foreground stars and background galaxies by their colors,
magnitude, redshift, proper motion, etc.

Yes, that is my idea as information going into this process. I've
been under the impression that Sloan and or Hipparchus contains a
listing of all point sources visible in medium-sized telescopes,
with enough data about each to make such a classification.

Then semi-accurate info about the sky coordinates and scale and
orientation of some old image, plus modern image-processing
software to match features between the image and a sky map, ought
to produce a near-exact mapping between image pixel coordinates and
sky coordinates, whereby the pixel coordinates of each cataloged
object can be calculated. Then an algorithm can verify that the
point source is visible in the image, and remove by averaging it
down to the local context so that it shows as a gray smudge instead
of a bright spot. It might even be possible to gray-out the
diffraction spikes around the spot.

Whether or not this has actually been done to
make a pretty picture -- rather than just to categorize the objects
for scientific study -- I don't know,

I posted to find out if anybody knew if this had actually been
done, at least for some of the common images such as M31 that are
so very commonly posted for public view. (I also posted it as a
challenge to anyone with access to all the data (and understanding
how to use it), and access to suitable image-processing software
(ditto), in case they might say "AFAIK it was never been done
before, but after you posted I did it!!".)

From: Bluuuue Rajah Bluuuuue@Rajah.
It's a good question, because you forgot the part where it would
be useful for producing a high contrast image, uncontaminated by
"light pollution" from foreground stars.

Yeah, that was most of my idea: When we look at an image such as
M31, we see lots of bright spots of light. We can't tell from
looking at the image which are foreground stars having nothing to
do with M31 and which are bright stars or globular clusters within
M31. This makes it very difficult to fully understand what we are
viewing. If we see something that looks out of place, we have no
way to know whether it's a feature of M31 that we need to
understand, or just a foreground star that we ought to not be
seeing in the image except for our location within our own galaxy
with hoardes of stars interfering with our view of M31.

The answer to his question is "probably not," since techniques
probably don't exist for masking out more than one star.

Perhaps you misunderstood my idea. I'm talking about
post-processing of the already-collected (archival) digital image
to remove artifacts including foreground stars, not physically
placing a occluding disk in the focal plane when making a new
image. Surely if you know the locations of ten foreground stars
within the archival image, and you can process the image to
"eliminate" one of the stars, you can repeat the process for each
of the other stars just the same.

The average person would use Photoshop for that, but did you stop to
think that there will then be a big black disk on the picture of the
galaxy? That would be only slightly less annoying than the star that
was originally blocking your view. :P

If the OP is serious, he should make a catalog of all the stars
blocking the images of all the easy to find galaxies, to see if
there's a situation where the contrast of one nearby galaxy could
be boosted by eliminating just a single star.

That's not my purpose. I expect removing the foreground stars from
an image won't significantly reduce the peak pixel intensity, thus
won't permit ramping up the contrast much more than it was set
already to achieve maximum span from black to white. Maybe some
single foreground star is brighter than the peak central brightness
of the core of M31, maybe not.

Also, I don't know how to get access to the key data I would need
to determine how many known foreground stars are visible in some
region around M31.

This should be possible by adapting the existing techniques. If
so, this idea is definitely worth a proposal for preliminary study,
and the OP could suddenly find himself a PI and an employed
scientist, rather than just a speculative amateur, sitting at his
PC and inventing wild ideas.

I really don't think this idea would warrant a professional
research project. But it might be useful for some company that
manufactures telescopes and sells a package of image processing
software and GoTo control for the telescope. It could brag that the
GoTo database includes all of the catalogs that exist to date, and
that it can automatically find all known objects within any region
of interest, and can edit images according to several algorithms
using that data, such as masking out foreground stars to gray (to
make them go away) or purple (to make them stand out like "sore
thumbs"). The advertisement could then show comparative images of
M31 with and without masking out of foreground stars. It could brag
"this is how M31 normally looks when viewed from Earth, and this is
how M31 *really* looks without our own galaxy blocking the view".

There is a programming language associated with Photoshop, but if you
want to do it yourself, you'd have to write a program to edit the image
file. It should be simple enough to just set the various pixels within
a specific range to black, but different photos would have different
exposures and contrasts, and I have no idea how you'd handle that.

Another possible type of person who might be interested in my idea
are the people who submit their best images to the Gallery section
of Sky and Telescope. I'd like to see somebody post an image of M31
with foreground stars removed. If S&T can post images of landscapes
at night with polar star trails in background, and images of an
eclipse viewed through pretty clouds, and images of sun dogs, I
think there's also room to post a galaxy image with foreground
stars removed. Once such an image gets constructed and published,
then maybe one of the advertisers would adopt the same idea in
their ads.

OT: Another neat thing I'd like to see is a stereoscopic view of
common asterisms such as the coathanger which have been proven as
chance line-ups of unrelated stars of varying distances. Or maybe a
3-stereoscopic view, with a central image just the way the asterism
looks from Earth, and views to each side showing how the asterism
would look as viewed a few lightyears to each side, so you can
choose either two adjacent images to get a stereoscopic view from
that side of the Earth line of sight. Or maybe a hologram that
shows true (scaled) distances which can be viewed from any
direction with any orientation.

The difference between a professional and an amateur is not just
having a good idea. It's having a good idea that can be made
into a real research project that get published. This could
probably be done, but it would take work. Maybe there's a needy
PhD student out there looking for a dissertation project.

Ignoring the fact that I don't think my idea would have any
significant true scientific value, only advertising value or
show-off value or public-education value (helping the public to
understand what a galaxy really looks like, without the distraction
of the foreground stars giving a wrong impression of the galaxy),
there's the fact that I don't have the English writing skill to
write a research or R&D proposal even for the kinds of projects
that I truly feel would be worth pursuing for scientific or
technological or society-improvement value. For example, see my Web
page listing several projects in Web software that I'd like to find
somebody else to work with me to brainstorm the design and help
with testing even if I do all the actual work.
http://shell.rawbw.com/~rem/WAP/projectIdeas.html
I have no idea how to find anyone interested in working with me on
any of those projects, and I have no idea how to write up any of
them as a formal proposal that anyone would pay serious attention
to.

I have a friend with a company that does precisely this, but he charges
real money. I'm not sure that he's ever written a proposal for a
government grant, either, so he pobably wouldn't be able to do what you
need. See if your phone book has listings for either "business plans"
or "business consultants."

And even if I did take a chance at writing a formal proposal, I
have no idea how to find anybody else to proofread it to show me
all the stupid mistakes I made so that I can make the proposal
halfway decent before I submit it to some funding agency.

From what I hear, it takes a few tries, but they will give you copies of
other people's grants that were approved.

From: Bluuuue Rajah Bluuuuue@Rajah.
The average person would use Photoshop for that,

Does the *average* person even *have* Photoshop in the first place?
It's my understanding that it doesn't come with any of the popular
operating systems, and is rather expensive to purchase, and illegal
to download without purchase. Is that correct? If so, I would guess
that a very small fraction of people with computers have Photoshop.
I don't know anyone around here who has Photoshop, and I'm pretty
sure none of the public computer labs around here have it. There's
just one person I knew in late 2000 who did a lot of editing of
photos and probably had it, but she lives 1.5 hours from here and
hasn't answered any of my e-mail for several years so I don't think
I can ask her to let me use her copy even if she has it.

but did you stop to think that there will then be a big black
disk on the picture of the galaxy?

I seriously doubt that would be the result. When I look at an image
of the Andromeda galaxy, I see some continous texture such as
spiral arms and hydrogen-alpha regions and the central bulge, and I
see several point sources, which I don't know whether they are
super-bright objects in the Andromeda galaxy or medium-bright
foreground stars in our own galaxy. If every last one of those
bright point sources are in fact blurred out, there wouldn't be
what you claim.

There is a programming language associated with Photoshop,

I've never even *seen* Photoshop, much less used it, so that's news
to me, good news if true. So it may be possible to use an external
program to automatically scour Sloan and/or Hipparcos databases to
produce a list of locations of foreground stars in the field of
view, and write out their approximate pixel coordinates. Then have
a Photoshop program import that data and match the given
coordinates with point-like bright spots in the actual image and do
least-squares fit between predicted pixel coordinates and actual
image bright-spot coordinates, and thereby produce *true*
foreground-bright-spot pixel coordinates. Then for each spot
calculate a peak brightness and radius of diffraction/blur based on
picture data, then do a least-squares fit of that info across all
the image to learn the effective diffraction size and blur function
as they vary across the image and to learn the response function
that relates true apparent magnitude with the brightness in this
image, in each of the color channels separately. Then use that
result to do the actual blurring wherever the predicted total
brightness-addition due to blurred diffracted star image is greater
than some threshold.

but if you want to do it yourself, you'd have to write a program
to edit the image file.

I've been looking for software libraries in any of the various
common programming languages that are able to process JPEG files,
for a different purpose but applicable for this task too, and
posted a query to comp.lang.lisp a few days ago, but nobody
responded yet with info I can use easily, according to Google
Groups which might not be working properly so I can't really be
sure nobody responded. Checked just now, found the thread again:
http://groups.google.com/group/comp....ex/browse_frm/
thread/85ee2ccfc158a0f5/d36d76676bfc469e?hl=en&_done=%2Fgroup%2
Fcomp.lang.lisp%2Fbrowse_frm%2Fthread%2F85ee2ccfc1 58a0f5%2Fd36d
76676bfc469e%3Fhl%3Den%26tvc%3D1%26q%3Dinsubject%2 53Ajpeg%2Baut
hor%253Auh3t%26&hl=en&tvc=1&q=insubject%3Ajpeg+aut hor%3Auh3t
= http://tinyurl.com/6oaur6
It has one reply I hadn't seen yet, but it isn't of much further
help, so I'm still lacking anything I can readily use.

It should be simple enough to just set the various pixels within
a specific range to black,

Ug! I'd rather blur than blacken. Ideally I'd have the ability to
create a flashing image, which flashes back and forth between
original and blurred, or back and forth between smooth-blurred and
checkerboard-dither-blurred, so that I can easily see which places
to ignore where I'm not seeing the true data in those regions.

Still, blackening the vicinity of foreground stars might be a quick
first step just to see if that's good enough so that I don't need
to do what I currently think I want to do.

but different photos would have different exposures and
contrasts, and I have no idea how you'd handle that.

Also different diffraction limit, quad or bi or hex diffraction
spikes depending on telescope design, different blurring due to
seeing conditions combined with exposure time, different function
change over whole image due to different coma correction or
astigmatism or spherical abheration or local displacement due to
variation of refraction from average seeing conditions due to
too-short exposure etc.

http://shell.rawbw.com/~rem/WAP/projectIdeas.html
I have no idea how to find anyone interested in working with me on
any of those projects, and I have no idea how to write up any of
them as a formal proposal that anyone would pay serious attention
to.
I have a friend with a company that does precisely this, but he
charges real money.

I have no real money whatsoever. It'll take me until I'm 80 years
old before I have credit cards paid off, if I live that long,
before I might possibly have income not already owed to credit
cards that I would be free to spend on anything for myself.

See if your phone book has listings for either "business plans"
or "business consultants."

If they pay money to get listed in the Yellow Pages, then
presumably the expect to charge me money I don't have to reimburse
them for their investment in Yellow Pages advertising.

"Robert Maas, http://tinyurl.com/uh3t"
wrote in message ...
From: Bluuuue Rajah Bluuuuue@Rajah.
The average person would use Photoshop for that,

Does the *average* person even *have* Photoshop in the first place?
It's my understanding that it doesn't come with any of the popular
operating systems, and is rather expensive to purchase, and illegal
to download without purchase.

I have an early version of Photoshop, a version that was issued free
with a scanner I purchased. Some functions are disabled.
Is the *average* person even *have* honesty, or will the *average*
person record music for his own personal use and think nothing of it?
You are not honest, that's for sure. You are probably average.

In article ,
(Robert Maas, http://tinyurl.com/uh3t) writes:
When we look at the Andromeda Galaxy, we see a mix of light from
the galaxy and vicinity itself, and light from foreground objects
such as stars within our own galaxy. To what degree have these
sources of light been classified as to which galaxy they belong to?

Probably not very well. In general, the brighter objects will belong
to the Milky Way and the fainter ones to M31, but there's a broad
range where they overlap. Taking spectra of all of them (to measure
"photometric parallaxes") would be time consuming and not very
interesting.

--
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA
(Please email your reply if you want to be sure I see it; include a
valid Reply-To address to receive an acknowledgement. Commercial
email may be sent to your ISP.)