DM has no formal requirements on how well we will process crowded fields; this needs to be fixed.
This is clearly a DRP science responsibility, but the DRP Scientist has no experience in this domain, so help/advice would be very welcome (I know at least Mario Juric, Zeljko Ivezic, Colin Slater, and Robert Lupton have more experience with this than I do).
Here's my rough plan for how to start:
- Find some examples of data at different galactic latitudes LSST, spanning the space of what LSST will observe (or the equivalent in stellar density due to other substructure). Ideally this would be HSC, DECam, or CFHT data.
- See how well DM processing currently handles these datasets.
- See how well a specialized crowded-field code handles these datasets.
I'm hoping that will help to reveal the quantities we'll actually want to put requirements on.
I do not anticipate starting this work until May 1 at the earliest; the April sprint is already quite overbooked for me.