Identification Of The Orbital Angular Momentum Of Vortex Beams Under Atmospheric Turbulence

The vortex beam has a unique spiral phase wave front,carries orbital angular momentum(OAM),and has the characteristics of barrel intensity distribution,small dark spot size and propagation invariance.Vortex beam has a wide application prospect in optical micro-manipulation,optical microscopy,space optical communication,precision measurement,and quantum information technology.The phase singularity expands with the increase of OAM quantum number and the propagation property of beam becomes unstable.In order to amplify the capacity of optical communication and quantum information processing,vortex beams with fractional OAM have been proposed.When the fractional vortex beam is transmitted in free space,it is affected by atmospheric turbulenc(AT),and the accuracy of the traditional recognition method is greatly reduced.Therefore,we use machine learning method to extract,learn and identify the characteristics of the profile of fractional vortex beams transmitted in free space.We proposed and verified a hybrid fractional dual-mode vortex beam OAM detection method,which provide two controllable degrees of freedom:quantum number l and proportional parameter n.The residual convolution neural network model(Res Net)was used to learn and identify the fractional OAM.High accuracy of identifying the fractional OAM disturbed by AT was realized.The relationship between different OAM quantum number l,proportional parameter n,turbulence intensity C_n~2,transmission distance z and the recognition accuracy is studied.Results show that for different l and n,even the turbulence intensity is strong,the propagation distance is far,the recognition accuracy also achieve above 85%,which shows that machine learning method has a broad application prospect in the free space optical communication,help to expand cross application research between machine learning and vortex beams.