Diffractive Metasurface‐Enabled Single‐Shot Characterization of Hybrid‐Order Poincaré Beams

Abstract Hybrid‐order Poincaré beams (HyOPBs) with complex transverse polarization states hold significant potential in optical communication and quantum information. Fully characterizing the Poincaré parameters of HyOPBs is a key task to accelerate applications. Conventional methods typically require mapping in multiple polarization states and reconstructing point by point, creating a fundamental bottleneck for fast measurement and real‐time monitoring of Poincaré parameters. In this work, a single‐shot and real‐time characterization scheme for HyOPBs is demonstrated by applying diffractive neural networks to a system of cascaded diffractive metasurfaces. The designed diffractive metasurfaces essentially function as an optical processor, efficiently extracting high‐dimensional spatial modes and complex amplitude information. Whereafter, the Poincaré parameters are accurately predicted, and the HyOPBs are correctly reconstructed with the help of electronic deep neural networks. This innovative approach is validated through a series of simulation studies with average reconstruction errors of <2.68% for σ and 1.84% for θ, respectively. The work provides an effective strategy for precise, compact, and real‐time detection of HyOPBs, paving the way for their application in the next generation of high‐capacity optical communication systems.