[vsnet-chat 5921] Re: ResSkySurvey and PIXY

[Brian Skiff <Brian.Skiff@lowell.edu>]

     Highly-automated surveys can be quite reliable as long as various
internal consistency checks can be included in the reduction sequence.
You can't simply get numbers out and accept them at face value.
     For our near-Earth asteroid survey at Lowell, for instance, we compute
notional orbits for each moving-object detection.  Those failing to fit
any sort of main-belt asteroid motion are examined on-screen by eye.
It turns out that extremely few of the false detections have main-belt
asteroid motion, so that ordinary main-belt asteroids need not be examined.
     For a photometric survey, among the checks one would like to see
include measuring photometric standard stars in each field as "unknowns"
to see how they match up against 'truth'.  If I were planning such a survey
(I would indeed like to!), among the strategies I would use (or consider)
is to take overlapping frames.  This could be done as sequential exposures
with a slight position offset (a few pixels is enough), or to have the
camera take pictures in a sequence that provided center-to-corner overlap
in a fairly short amount of time ( < ~1 hour).  Another tack is to have twin
cameras or expose in two filters (V and I or whatever).  For variables with
relatively long cycles, intra-night magnitudes had better be highly
consistent, as should colors and coordinates if measured.  Again, the local
standard stars not observed as part of the photometric solution provide an
internal check on quality.  If sets of observations fall outside some limit
determined by these, then a flag gets set to indicate some problem.
     Surely the TASS, ASAS, etc. survey folks have thought about this and
could provide some relevant advice in terms of what tests can be applied
and how to automate them.  Folks shouldn't have to reinvent the wheel here.

\Brian

VSNET Home Page

Return to Daisaku Nogami

---

vsnet-adm@kusastro.kyoto-u.ac.jp