search

michaelaye / planet4

Analysis software for the PlanetFour citizen science project.
www.planetfour.org
ISC License
2 stars 3 forks source link

Standard clustering eps value of 10 too harsh to catch disperse areas #22

Open michaelaye opened 8 years ago

michaelaye commented 8 years ago

When the deposits are dispersed widely without showing fan or blotch shapes, the standard criterion of 10 pix to match markings into a cluster are too harsh, as can be seen in APF000010p

I'm worried if I loosen this criterion that more false positives are being created, but this remains to be studied.

michaelaye commented 8 years ago

apf000010p_pipeline_check

michaelaye commented 8 years ago

Check out APF0000185 for this as well.