2005MNRAS.358..883K

We investigate blending, binarity and photometric biases in crowded-field CCD imaging. For this, we consider random blend losses, which correspond to the total number of stars left undetected in unresolved blends. We present a simple formula to estimate blend losses, which can be converted to apparent magnitude biases using the luminosity function (LF) of the analysed sample. Because of the assumptions used, our results give lower limits of the total bias and we show that in some cases even these limits point towards significant limitations in measuring apparent brightnesses of standard candle stars and thus distances to nearby galaxies. A special application is presented for the OGLE-II BVI maps of the Large Magellanic Cloud (LMC). We find a previously neglected systematic bias up to 0.2-0.3 min for faint stars (V∼ 18.0 -19.0 min) in the OGLE-II sample, which affects LMC distance measurements using RR Lyrae and red clump stars. We also consider the effects of intrinsic stellar correlations, i.e. binarity, via calculating two-point correlation functions for stellar fields around seven recently exploded classical novae. In two cases, for V1494 Aql and V705 Cas, the reported close optical companions seem to be physically correlated with the cataclysmic systems. Finally, we find significant blend frequencies up to 50-60 per cent in the samples of wide-field exoplanetary surveys, which suggests that blending calculations are highly advisable to be included into the regular reduction procedure.