Hubble GSC Image
Ernie Wright
I've created an image of the Milky Way by plotting the stars in the
Hubble Guide Star Catalog, and I'm hoping to get some help understanding
the tile artifacts in the image.
http://www.access.digex.net/~erniew/images/gsc.jpg (141 KB)
The catalog was derived from a lot of separate plates, of course, but
the tiles in the image cover a larger area than the plates. For those
with the patience to load a second image,
http://www.access.digex.net/~erniew/images/gscp.jpg (159 KB)
overlays what I think are the plate boundaries. (The boundaries emerged
when I allowed stars with only one plate number to affect each pixel.)
So what's the deal with the tiles?
- Ernie http://www.access.digex.net/~erniew
Brian Skiff
In article <3466AE...@access.digex.net>, Ernie says...
>
>Brian Skiff wrote:
>
>> One more reason to be careful when you use GSC data.
>
>Thanks for the reply, Brian. Your explanation of the edge effects makes
>sense, and I'm sure you're right that I'd confused plates with regions.
>I'm working with a cut down version of the GSC and wasn't aware of the
>distinction.
>
>I'm trying the same kind of plot with the Tycho database (obtained on CD
>from the ADC, http://adc.gsfc.nasa.gov). So far it looks like the tiling
>will be a lot less apparent, but there's also less detail.
>
>- Ernie http://www.access.digex.net/~erniew
>
The "perfect" star catalogue is not yet in hand, although a combination
of the "Astrographic Catalogue" re-reduction (called AC2000) and the
re-reduction of the GSC (GSC version 1.2) would be very useful for a lot of
tasks. You can read about AC2000 here:
http://aries.usno.navy.mil/ad/ac.html
A brief description of GSC v1.2 with a link to the look-up utility, can be
found here:
http://www-gsss.stsci.edu/gsc/gsc.html
As an aside on the appearance of your plots, it is interesting to see how
the various plates series used to build the catalogue affect the appearance
of the sky. Specifically, the southern Milky Way does not have nearly as many
stars as it should if the GSC were complete to the same limits as in the north.
The reason for this is that in most areas blue-light plates were scanned,
resulting in magnitudes fainter than they would be if orange-light plates were
used, as in the north. Also right along the southern galactic plane, in order
to avoid problems with crowding, a series of very-short exposure plates was
used, reducing the apparent star density even further. The STScI folks only
wanted enough stars for guiding HST and didn't care about getting stars to some
magnitude limit---they didn't have to go very faint in Carina to have enough
stars!
These points are merely manifestations of why I've been telling people in
discussions about the GSC on this newsgroup that the GSC _is_not_ an inventory
of the sky to some limit, but merely an engineering product for pointing and
tracking Space Telescope.
\Brian Skiff (Lowell Observatory)
Brian Skiff
In article <346591...@access.digex.net>, Ernie says...
These are very interesting plots, and I'm glad you've made them available.
At first blush, just looking at familiar patches of sky, it seems to me as
though the second plot showing the nominal plate boundaries has the squares
too small. They look instead like the "small regions" in the GSC (i.e. the
first four digits of the GSC name), which IIRC are roughly (but not always)
about 3 degrees square.
The "frames" come about because when the catalogue was compiled, they
used an algorithm to match stars with similar positions so that objects within
about 2 arcsec of each other were given the same name. If you've looked in
the GSC you know that in the plate-overlap regions there are often two or
three (even four and five) entries for the same star with very slightly
different positions. The problem is that on the edges/corners of the plates
the positions were bad enough that this matching algorithm broke down, since
positions for the same star were several arcseconds apart, so got included as
separate entries in the catalogue, even though there's only one star at that
place.
So the outlines of the plate boundaries arise because there are
_apparently_ more stars around the edges of the plates. In fact, because of
vignetting on the Schmidt plates, there are actually fewer stars recorded
there, but because of the way the catalogue was built, these rather large
astrometric errors remain, producing spurious duplicate entries.
One more reason to be careful when you use GSC data.
\Brian
Ernie Wright
Brian Skiff wrote:
> One more reason to be careful when you use GSC data.
Thanks for the reply, Brian. Your explanation of the edge effects makes
Chris Marriott
In article <3466AE...@access.digex.net>, Ernie Wright
<ern...@access.digex.net> writes
>Thanks for the reply, Brian. Your explanation of the edge effects makes
>sense, and I'm sure you're right that I'd confused plates with regions.
>I'm working with a cut down version of the GSC and wasn't aware of the
>distinction.
>
Another point to note is that the limiting magnitude differs between GSC
regions, based on the "star density" in that part of the sky. If you
display GSC stars down to mag 15, say, you'll often see a sharp "edge"
between regions where the database's limiting mag changes.
Chris
----------------------------------------------------------------
Chris Marriott, Microsoft Certified Solution Developer.
SkyMap Software, U.K. e-mail: ch...@skymap.com
Visit our web site at http://www.skymap.com
Ernie Wright
It probably doesn't matter much, but after working with the Tycho data,
I've decided that my GSC plots were upside-down and off-center. I've
posted revised ones.
http://www.access.digex.net/~erniew/images/gsc.jpg (141 KB)
http://www.access.digex.net/~erniew/images/gscp.jpg (160 KB)
- Ernie http://www.access.digex.net/~erniew
Ernie Wright
Brian Skiff wrote:
> http://www-gsss.stsci.edu/gsc/gsc.html
I was especially interested in the summary of the systematic errors and
the efforts made to correct them in version 1.2,
http://www-gsss.stsci.edu/gsc/gsc12/description.html
although as a qualitative illustration, I kinda like my image better than
the figures there.