Restoration of astrophysical images, statistical influence of the camera
The restoration of astrophysical images is a classical application of inverse problems. The astronomical object is first blurred by the Point Spread Function of the instrument-atmosphere set. The convolved image resulting is corrupted by a Poissonnian noise in reason to low light intensity, then a Gaussian white noise is added during the electronic read-out operation by the Charge Coupled Device (CCD) camera, leading to a "Poisson Gaussian" density. Two basic cases with positivity constraints have been fist considered, either a pure Poisson noise leading to the so-called Richardson Lucy algorithm or a pure Gaussian additive noise corresponding mainly to infrared experiments with a high intensity level, leading to the so-called ISRA algorithm. The complete model corresponding to the Poisson Gaussian process has been also developed. More recent technology proposes to acquire astrophysic data with Low Light Level CCD (L3CCD) cameras in order to avoid the read-out noise due to the classical CCD acquisition. The physical process leading to the data has been previously described by a "Poisson Gamma" density. We propose to discuss the statistical models in the CCD and L3CCD cases and an illustration is given on synthetic astrophysical data chosen to discuss the interest of one over the other according to the considered experiments.