Coded mask imaging of extended sources with Gaussian random fields

Abstract

A novel method for generating coded mask patterns based on Gaussian random fields (GRF) is proposed. In contrast to traditional algorithms based on cyclic difference sets, it is possible to construct mask patterns that encode a predefined point spread function (PSF). The viability of this approach and the reproducibility of the PSFs is examined, together with studies on the mean transparency, pixel-to-pixel variance and PSF deterioration due to partial shadowing. Sensitivity considerations suggest the construction of thresholded realisations of Gaussian random fields (TGRF) which were subjected to the same analyses. Special emphasis is given to ray-tracing simulations of the pattern's performance under finite photon statistics in the observation of point sources as well as of extended sources in comparison to random masks and the pattern employed in the wide field imager onboard BeppoSAX. A key result is that in contrast to traditional mask generation schemes, coded masks based on GRFs are able to identify extended sources at accessible photon statistics. Apart from simulating on-axis observations with varying levels of signal and background photon counts, partial shadowing of the mask pattern in the case of off-axis observations and the corresponding field-of-view is assessed.