Keywords: DataAccWG
Attendees
Tim Axelrod
Serge Monkevitz
Deborah Levine
Robert Lupton
Kem Cook
Ani Thakar
Maria Nieto-Santisteban
Jacek Becla
confusion / one object becoming two...
=======================================
During deep detection we optimally process all available
information simultaneously, so during that process we know
everything about the confusion
SDSS approach
- match table: matches fussy things
- still in research
During nightly processing we will deal with DIA images,
which will be relatively non-crowded, and reasonably clean
- still, there are some nasty cases, for example asteroid
moving past star
Database has to support many-to-many relationship between
DIASources and Objects, unless we can choose on the fly
during processing DIA image how to associate DIASource
with Object (1-to-1)
Let's suppose we could run deep detection every night,
How would we solve the many-2-many mapping?
In steady state (after at least one deep detection),
we will have most objects resolved, that helps...
Suppose we see 1 "blob" instead of expected 3 stars, how
do we assign this measurement to these 3 stars? Do we
partition the flux? Do we use the best fit to assign
the measurement to one star only?
We should think first about steady processing, and don't
worry too much about the bootstrapping the database
(the very first few nights/weeks)
- perhaps we could leverage commissioning (the test
runs before the production operation starts) and use
that to bootstrap the database?
We will come back to this topic, it definitely needs
more exploring
- Deborah will try to put together a cartoon/some notes
- we will all try to document our thinking/current knowledge
- it should definitely address issues like
- what if one source matches two objects?
- what if two sources match one object?
managing schema
===============
will keep master schema in EA, and generate ascii version
from that
- existing approach (master in ascii file) error prone and messy
USNO
====
have the USNO data on SLAC server now
AOB
===
Next big meeting Jan 12
Thanks,
Jacek
Jacek Becla wrote:
Keywords: DataAccWG
Reminder: we will have a 2-hour DataAccWG telecon tomorrow,
Friday starting at 9:00 PST
call-in ip address for video: 140.252.1.34
dial-in: 866 330 1200 passcode 300 2363
Agenda
- "one object becoming two" issue...
- other topics related to nightly processing use case as time permits
Jacek
_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data
_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data
