Mario Ivanov (Royal Military College of Canada), Donald McGaughey (Royal Military College of Canada)
Keywords: Phase diversity, aperture diversity
Abstract:
Phase diversity is a well known method to estimate a phase aberration and undistorted object. In this paper we present a new method of collecting diverse images called aperture diversity. This technique involves forming images through different shaped apertures. Incoming wavefronts are collimated and beam-split into two or more paths to ensure the wavefront in each path have the same aberrations. Images are then formed for each path using apertures with different sizes. Since the optical transfer function of the system is an autocorrelation of the generalized pupil function, the transfer functions for each path are different but related, thus giving a strong constraint for minimization. We use two circular apertures – the first being un-obscured, the second having a central circular obscuration. The radius of the central obscuration varies in different experiments but typical values are between 0.4 and 0.7 of the whole aperture. Blurred images are simulated following the Kolmogorov model of atmospheric turbulence. For each frameset an independent phase screen is constructed. The pupil phase is calculated for each of the two channels and two sets of simulated images are produced. Different magnitudes of the simulated turbulence are investigated, ranging from mild to severe. A combination of two minimizing algorithms is used to obtain the set of Zernike weight estimates – the Simplex algorithm and Genetic algorithm. The weight estimates are then used to construct object intensity estimates using a Wiener filter. The object estimates are compared visually and quantitatively to the initial undisturbed object and the blurred input images. Different radii of the central obscuration are investigated in searching of the best results. The method is described and initial results for real images collected with an imaging system and a static phase screen are presented.
Date of Conference: September 12-15, 2007
Track: Poster