Polarimeter blind deconvolution using image diversity

This research presents an algorithm that improves the ability to view objects using an electro-optical imaging system with at least one polarization sensitive channel in addition to the primary channel.; Following the review of historical methodologies applicable to this research area, the statistical Cramer-Rao lower bound (CRLB) is developed for a two-channel polarimeter. The CRLB is developed using the system's ability to resolve two point sources in the presence of atmospheric turbulence. The bounds show that such a polarimeter has an advantage over previous imaging methods at smaller separations.; A small optical laboratory is set up to generate a set of calibrated images for verification of the simulation results and validation of algorithm development. Defocus is the aberration chosen for algorithm development and testing due to its significant presence when imaging through turbulence and its ease of production in the laboratory. An innovative algorithm for detection and estimation of the defocus aberration present in an image is also developed.; Using a known defocus aberration, an iterative polarimeter deconvolution algorithm is developed using a generalized expectation-maximization (GEM) model that produces results as predicted by the CRLB results. Using an example bar target set with a degree of polarization of one, the polarimeter deconvolution algorithm can resolve the two bars down to half the bar separation as the Richardson-Lucy (RL) algorithm can do. In addition, a fidelity metric is used that shows the polarimeter deconvolution algorithm deconvolves simulated targets with approximately half of the error present in objects deconvolved using the RL algorithm.; The polarimeter deconvolution algorithm is extended to an iterative polarimeter multiframe blind deconvolution (PMFBD) algorithm with an unknown aberration. Using both simulated and laboratory images, the results of the new PMFBD algorithm clearly outperforms an RL-based MFBD algorithm. The convergence rate is significantly faster with better fidelity of reproduction of the targets.; This research successfully developed an algorithm that uses polarization data in conjunction with standard imaging to improve the spatial resolution of deconvolved objects with faster convergence rates. Clearly, leveraging polarization data in electro-optical imaging systems has the potential to significantly improve the ability to resolve objects and, thus, improve Space Situation Awareness.