search

coolneighbors / unWISE-verse

Integrated flipbook / panoptes pipeline
MIT License
2 stars 0 forks source link

How to derive the best min/max pixel values for a given region #17

Closed austinh2001 closed 2 years ago

austinh2001 commented 2 years ago

Dan Caselden brought up a valid and important question in Slack about this. Merits future discussion.

austinh2001 commented 2 years ago

Aaron Meisner on Slack: "my usual/default thing is linear stretch from 2.5th percentile to 97.5th percentile in terms of distribution of pixel values"

EdenSchapera commented 2 years ago

Plan going forward -- run a "poll" on zooniverse 3 sets (-30 to 120) (-20 to 120) (-10 to 120) (10 targets per set, 2-3 known brown dwarfs per set, retirement of 3) to see which min/max value ranges allow the testers to be the most accurate.

It would be nice to have the original pixel flux values, but otherwise we can just optimize for user experience and accuracy.

austinh2001 commented 2 years ago

Currently trying to do this process with unWISE cutouts directly requested from unWISE rather than requesting via WiseView

EdenSchapera commented 2 years ago

Closed with recent changes