headers
Tim Axelrod | 31 Jan 2008 15:51
Picon

One parameter tweak prior to final DC2 runs

Hi Ray,

I have one parameter tweak for detection that I'd like to put in place 
before our final DC2 runs, to numPixMinFootprint.   It's very clear that 
I have this set too small for the CFHT/Megacam data, and this is the 
primary cause of the large number of noise sources coming out of 
detection.   Yesterday's change to the detection pipeline was to store 
the number of pixels in each Footprint in the database entry for its 
corresponding DIASource.   A DC2 run over a few visits will let me 
gather the data to make an optimal choice for numPixMinFootprint.   Then 
we should be ready for the final runs.   Depending on how things go at 
your end, I'd hope to be done with that today.

Cheers,
Tim
Attachment (taxelrod.vcf): text/x-vcard, 259 bytes
_______________________________________________
LSST-data mailing list
LSST-data <at> lsstcorp.org
http://listserv.lsstcorp.org/mailman/listinfo/lsst-data
Permalink | Reply |
headers
Ray Plante | 31 Jan 2008 16:02
Picon

Re: One parameter tweak prior to final DC2 runs

Hi Tim,

The event broker locked up last night, so I wasn't able to do any runs 
over night, so go ahead and make the changes.  Ping me when you are done.

Is this a change in code or policy?

thanks,
Ray
Permalink | Reply |
headers
Tim Axelrod | 31 Jan 2008 16:42
Picon

Re: One parameter tweak prior to final DC2 runs

Hi Ray,

This will only be a policy change.   I want to make the change only 
after seeing the results of a run with the current settings.   So go 
ahead and do one whenever you're ready.

Cheers,
Tim

Ray Plante wrote:

>Hi Tim,
>
>The event broker locked up last night, so I wasn't able to do any runs 
>over night, so go ahead and make the changes.  Ping me when you are done.
>
>Is this a change in code or policy?
>
>thanks,
>Ray
>_______________________________________________
>LSST-data mailing list
>LSST-data <at> lsstcorp.org
>http://listserv.lsstcorp.org/mailman/listinfo/lsst-data
>  
>
Attachment (taxelrod.vcf): text/x-vcard, 295 bytes
_______________________________________________
(Continue reading)

Permalink | Reply |

Gmane