RE: Minimum spanning circles and Voronoi diagrams

[Tom Droege <tdroege2@earthlink.net>]

When we have experts advising us that have struggled with such problems for 
years, my inclination is to do what they say unless we can think of a very 
good reason why we are somehow different.  Arne's arguments make great 
sense to me.  Use the high precision UCAC measurements as the 
standard.  When there are two within our detection accuracy, mark as a 
blend.  Seems to me there are three cases:

1)  No bright stars within n" or so.  A clear separate star for tass
2)  Two stars with modest magnitude difference closer than n" but not 
closer than m"  This star is always marked as a blend.  We can expect big 
jumps in apparent magnitude.
3)  Two stars with modest magnitude difference closer than m".  These can 
never be separated by tass and so are counted as one star.

Working with a real data set will generate values for n and m.  I have a 
data set available to be used for such a purpose.

The result of all this is that we must have some flags carried with our 
data that mark such situations.

Tom Droege

At 03:43 PM 2/7/02 -0700, you wrote:
>Chris wrote:
> >what's your suggextion for dealing with doubles in UCAC
> >that are not resolved by TASS?  Do we pick one of the pair
> >and asign all observations to it. Asigning the observation to
> >the closest match will have kind of a random ping pong
> >effect.  Eventually we will have to look for these and
> >create a new catalog object to stand for the unresolved
> >double.  It gets harder when it resolves in only one
> >color or b 2 TASS sites but not the others.  I agree
> >catalog seeding is a good idea but it can't solve these
> >types of problems.
>   Basically this is a matching problem, not a seeding problem.
>As Dirk mentioned, the proper way to handle this is if more
>than one star in the UCAC catalog falls within the matching
>range of the imported coordinates, then identify the imported
>object with, say, the brightest UCAC object,
>but flag it as 'unresolved' or 'blended'.  Take UCAC as the
>Bible; it will have every normal object that a TASS camera can
>detect.  There will be a few new objects such as novae and
>asteroids, but they will be a very small subset.
>Arne