Topological charge recognition of vortex beams based on a convolutional neural network

The application of vortex beams in optical communications is significantly constrained by wavefront distortions caused by atmospheric turbulence and system-induced aberrations. In high-energy laser systems, spherical and coma aberrations are commonly introduced due to thermal effects during manufacturing or beam propagation. To address the challenge of topological charge identification under such complex distortions, this paper proposes an enhanced IResNet18 model capable of simultaneously predicting the topological charge, spherical aberration coefficient, and coma aberration coefficient from distorted intensity patterns. Three experimental configurations are designed: vortex beams with spherical aberration, vortex beams with coma aberration, and vortex beams with combined spherical and coma aberrations. This study systematically investigates the impact of varying propagation distances and turbulence intensities on model performance. Results demonstrate that the proposed model achieves superior accuracy and training efficiency compared to existing models. These findings provide a robust framework for aberration-aware vortex beam recognition and offer valuable insights for enhancing the reliability of free-space optical communication systems.