Summary: A general Maximum-Likelihood algorithm is described for fitting arbitrary models to CCD samples, where data are modelled as a Poisson process plus (gaussian) readout noise. It can be applied for example to 1D (LSF) and 2D (PSF) centroiding, but also for estimating the parameters of the LSF, PSF or CDM models.
 

Bibtex entry for this abstract:

@UNPUBLISHED{LL:LL-078,
author = {L.~Lindegren},
title={{A} general {M}aximum-{L}ikelihood algorithm for model fitting to {C}{C}{D} sample data},
institution={Lund Observatory},
year={2008},
month={November},
url={http://www.rssd.esa.int/doc_fetch.php?id=2861864},
note={GAIA-C3-TN-LU-LL-078},
type={Technical note}
}