Recognition of the polarization structure of synthesized vector beams in a turbulent atmosphere from intensity images by neural networks

In recent years, various approaches to using laser beams to increase the capacity of information transmission channels have been actively studied. In atmospheric optical communication channels, it is important to determine the characteristics of beams that can carry additional information when beams are distorted in the turbulent atmosphere. Coherent laser beam combining systems are promising for transmitting information encoded in the polarization structure of a laser beam, as they allow it to be changed at high frequencies. In addition to developing methods for encoding information with structured beam parameters, it is also necessary to develop methods for their receiving and decoding. Determining the polarization structure of a beam distorted by a turbulent atmosphere is a non-trivial task. The aim of this study is to investigate the possibility of using neural networks to determine the polarization structure of synthesized laser beams through intensity distribution images distorted by atmospheric turbulence. The study is based on numerical simulation. For the first time, it has been shown that the use of neural networks makes it possible to distinguish a linearly polarized synthesized beam from a beam with an inhomogeneous polarization distribution formed by sub-beams with azimuthally or radially distributed polarizations in a turbulent atmosphere.