Purity of galaxy sample better than 1% for i<22, even in crowded fields (R-31).
Specifically: Bright stars in dense fields shall not be confused with BCGs.
- Imaging to full depth at least in i-band of several dozens of clusters, preferably at low galactic latitude to increase the number of stars.
- For all clusters, spectroscopic truth information (star vs. galaxy) down to i = 22.
Possible fields: CFHTLS W1, where it overlaps with 2x2 deg VVDS shallow 0226 field.
- Ideally: cross check with HST imaging (single-band, only fairly short exposure data required) on several prominent clusters.
Todo: check HST imaging availability for these clusters: External
For all identified galaxies with i < 22, find those that are actually stars. Use spectroscopy or visual confirmation (HST imaging) to identify stars.
N* / N'_gal_22 < 0.01, where N* is the number of true stars among the objects identified as galaxies with i < 22 (whose number is N'_gal_22). This requirement is not probabilistic like e.g. R-28.
Improvements in either deblender or S/G required. The latter can be tested with SG-G2. Failure here with complicate to SG-G2 indicates problems of the deblender.