Image reconstructions with active illumination in strong-turbulence scenarios with single-frame blind deconvolution approaches.

Abstract

Image formation over long horizontal or slant paths is of interest in surveillance and remote sensing. Image reconstructions of isolated objects are presented using active illumination in long-path, strong-turbulence conditions using a wave-optics simulation to produce the images. Fast-running reconstruction algorithms are used, including a novel single-frame blind iterative deconvolution algorithm and a generalized expectation maximization algorithm. Significant improvements in image quality and image recognizability can be found for spherical-wave Rytov variances up to 0.4 and for up to 10 atmospheric coherence lengths across the aperture in uniform-turbulence scenarios over a 30km range. For these conditions, the isoplanatic patch angle is comparable to the diffraction angle, and there are 20 or more isoplanatic patches across the objects considered. The results are compared with idealized atmospheric phase correction with an incoherent beacon. Several image quality metrics are considered. Results for strongest turbulence are explained in terms of a local average of the point spread function and the central limit theorem for cases in which there are many isoplanatic patches across the object.