Improved prior for adaptive optics point spread function estimation from science images: Application for deconvolution

Context . Access to knowledge of the point spread function (PSF) adaptive optics(AO)-assisted observations is still a major limitation when processing AO data. This particularly important image analysis requires use deconvolution methods. As PSF complex and time-varying function, reference PSFs acquired on calibration stars before or after scientific observation can be too different from actual used for deconvolution, lead artefacts in final image. Aims We improved existing PSF-estimation method based so-called marginal approach by enhancing object prior order make it more robust suitable resolved extended objects. Methods Our process two-step blind literature. The first step consists estimation science For this, we made an analytical model, whose parameters are estimated algorithm. was then deconvolution. In this study, investigated requirements terms parameter obtain accurate yet resilient using simulations. show that current algorithms do not provide required level accuracy, especially presence small Therefore, modified algorithm providing new model description, leading parameters. Results fulfills requirement with realistic system configurations classes Solar System objects Finally, validate our performing SPHERE/ZIMPOL Kleopatra asteroid.