Siamese network-based vortex beam correction and identification in oceanic turbulence

Vortex beams carrying orbital angular momentum (OAM) increase the channel capacity of underwater wireless optical communication (UWOC) systems but are susceptible to oceanic turbulence (OT), resulting in the degradation of communication quality. Therefore, the correction of distorted vortex beams and the identification of OAM modes are crucial for UWOC. In this work, a joint distortion correction and OAM mode identification approach is proposed and experimentally verified based on a Siamese network (SN). The SN performs feature extraction and feature fusion on the phase screen and intensity pattern. In addition, a classification network is integrated into the SN framework, thereby enabling both mode identification and prediction of Zernike polynomial coefficients with a limited number of samples. The results show that the improved SN can quickly and accurately identify four OAM modes and predict the Zernike polynomial coefficients of four OT levels. The phase screen reconstructed with the predicted Zernike polynomial coefficients enables high-quality correction of the distorted vortex beam. The proposed SN-based vortex beam correction and identification exhibits robust generalization performance under small-sample conditions, opening a new path for UWOC.