Details
-
Type:
RFC
-
Status: Implemented
-
Resolution: Done
-
Component/s: DM
-
Labels:None
Description
Bright star masks
A set of pre-defined bright star masks are applied by AssembleCoaddTask. These are mainly intended as a convenience for the user: while it's possible for the pipeline to respect these masks, to my knowledge they are typically only used for setting pixel flags in the measurements. The current version of the bright star masks were created by Jean Coupon, and are known as the Arcturus masks (being the third version he created).
Andy Goulding (Princeton) has created a new set of masks that we believe are superior to the Arcturus masks because they are based on Gaia DR2, and have a more careful calibration of the mask size as a function of stellar magnitude. These masks will be used by the HSC Strategic Survey Program for the next production run.
Proposal
I propose installing these masks in the /datasets/hsc data repo on the LSST cluster, and making them the default. This will allow the processing by LSST to more closely match that by HSC moving forward.
Caveats
Use of these masks will require minor upgrades to the mask reading code, which I hope will merge soon.
Attachments
Issue Links
Activity
You're not going to want to use this mask for LSST, as it only covers the HSC survey area.
I don't expect the masks to change very often at the moment, but I agree some thought by the DM architects should be put into supporting multiple versions moving forward.
Paul Price
added a comment - You're not going to want to use this mask for LSST, as it only covers the HSC survey area.
I don't expect the masks to change very often at the moment, but I agree some thought by the DM architects should be put into supporting multiple versions moving forward. No dissent, adopted.
Paul Price
added a comment - No dissent, adopted. 
This sounds like a good idea. My only question is whether our bright star mask code needs to be made flexible enough to be able to deal with new mask databases in the future for LSST (a la the refObjLoaders), or whether such masks don't change often enough to warrant that.