Image Recovery

Abstract

The diffraction-limited phase retrieval of a degraded image is an important subject. Apart from the post-processing speckle interferometry, the adaptive optics systems may also require image-processing algorithms since the real-time corrected image is often partial. For a long baseline interferometry, the hypertelescope in particular, the image processing technique using bispectrum method (see Sect. 8.2.2) is needed. From the results of reconstructed image applying such a method for a (prototype) hypertelescope, where different aperture distributions and various objects were attempted, it appears that bispectrum algorithm is useful even before it becomes equipped with adaptive phasing. The natural aperture rotation, relative to celestial North, during hours of observing (possibly intermittent) may improve the reconstruction. Also, with a Carlina geometry, the pupil drift across a fixed and non-redundant array of mirrors tends to vary the aperture pattern, which improves the reconstructed image. Prior to using such algorithms, the basic operations to be performed are dead pixel removal, debiasing, flat fielding, sky or background emission subtraction, and suppression of correlated noise. The fringes are necessary to be examined and the visibility and contrast can be converted using Fourier transforms, so that the object under observation is mapped effectively. The visibility measured by such an instrument is characterized by the amplitude and phase of the fringe at different instants.