Crosstalk-Mitigated OAM Recovery Through Strong Scattering via Sparsity-Regularized LG-TM

Light beams carrying orbital angular momentum (OAM) provide an unbounded set of mutually orthogonal spatial modes, enabling high-dimensional information encoding for optical communication and sensing. However, strong scattering severely distorts the helical phase structure of OAM beams, introducing inter-modal crosstalk that fundamentally limits practical unit-interval OAM-encoded systems. Here, we introduce a sparsity-regularized framework for OAM recovery through strongly scattering media by exploiting the intrinsic sparse representation of OAM states in the Laguerre–Gaussian basis. Compared with the conventional Gerchberg–Saxton algorithm, the resulting sparsity-regularized framework improves both the accuracy and robustness of modal reconstruction, particularly under noisy and challenging scattering conditions. Experimentally, we faithfully recover OAM modes across (Formula presented.) at unit spacing with fidelities above 99% and a maximum crosstalk of (Formula presented.) dB, achieving a fourfold expansion of the usable OAM bandwidth and an approximately 12.6-fold reduction in linear crosstalk over previous methods. We further demonstrate error-free grayscale image transmission through strong scattering, together with high-fidelity reconstruction of complex, random, and Gaussian OAM spectra, highlighting the robustness and versatility of our scheme. The proposed design offers a scalable and practical pathway toward high-capacity communication systems by fully leveraging adjacent low-order OAM modes, benefiting both classical and quantum high-dimensional OAM-based communications, even under strong scattering conditions.