COMPENSATION FOR READOUT NOISE IN CCD IMAGES

Data acquired with a CCD camera are modeled as an additive Poisson-Gau ssian mixture, with the Poisson component representing cumulative coun ts of object-dependent photoelectrons, object-independent photoelectro ns, bias electrons, and thermoelectrons and the Gaussian component rep resenting readout noise. Two methods are examined for compensating for readout noise. One method relies on approximating the Gaussian readou t noise by a Poisson noise and then using a modified Richardson-Lucy a lgorithm to effect the compensation. This method has been used for res toring images acquired with CCD's in the original Wide-Field/Planetary Camera aboard the Hubble Space Telescope. The second method directly uses the expectation-maximization algorithm derived for the Poisson-Ga ussian mixture data. This requires the determination of the conditiona l-mean estimate of the Poisson component of the mixture, which is acco mplished by the evaluation of a nonlinear function of the data. The se cond method requires more computation than the first but is more accur ate mathematically and yields modest improvements in the quality of th e restorations, particularly for fainter objects. As a specific exampl e, we compare the two methods in restorations of images representative of those acquired with that camera; they contain excess blurring that is due to spherical aberration and a rms readout noise level of 13 el ectrons.