• Photonics Research
  • Vol. 9, Issue 4, B81 (2021)
Xiao Wang1, Yufeng Qian1, JingJing Zhang1, Guangdong Ma1, Shupeng Zhao1, RuiFeng Liu1、*, Hongrong Li1, Pei Zhang1, Hong Gao1, Feng Huang2、3, and Fuli Li1
Author Affiliations
  • 1Shaanxi Province Key Laboratory of Quantum Information and Quantum Optoelectronic Devices, School of Physics, Xi’an Jiaotong University, Xi’an 710049, China
  • 2School of Mechanical Engineering and Automation, Fuzhou University, Fuzhou 350108, China
  • 3e-mail: huangf@fzu.edu.cn
  • show less
    DOI: 10.1364/PRJ.412965 Cite this Article Set citation alerts
    Xiao Wang, Yufeng Qian, JingJing Zhang, Guangdong Ma, Shupeng Zhao, RuiFeng Liu, Hongrong Li, Pei Zhang, Hong Gao, Feng Huang, Fuli Li. Learning to recognize misaligned hyperfine orbital angular momentum modes[J]. Photonics Research, 2021, 9(4): B81 Copy Citation Text show less
    (a) Alignment-free fractal multipoint interferometer. Laser, He–Ne laser with 633 nm wavelength; L1, 50 mm lens; L2, 500 mm lens; SLM, phase-only spatial light modulator; L3, 300 mm lens; P, pinhole; L4, 300 mm lens; DMD, digital micromirror device; L5, 250 mm lens; CCD, charge-coupled device. (b) proposed FMM; (c) example of the far-field intensity patterns.
    Fig. 1. (a) Alignment-free fractal multipoint interferometer. Laser, He–Ne laser with 633 nm wavelength; L1, 50 mm lens; L2, 500 mm lens; SLM, phase-only spatial light modulator; L3, 300 mm lens; P, pinhole; L4, 300 mm lens; DMD, digital micromirror device; L5, 250 mm lens; CCD, charge-coupled device. (b) proposed FMM; (c) example of the far-field intensity patterns.
    Schematic diagram of DenseNet-121. CONV, convolution layer; MP, max pooling layer; DB, dense block; GMP, global max pooling layer; FC, fully connected layer.
    Fig. 2. Schematic diagram of DenseNet-121. CONV, convolution layer; MP, max pooling layer; DB, dense block; GMP, global max pooling layer; FC, fully connected layer.
    Examples of the experimental diffraction intensity patterns for LG eigenstates with topological charge ℓ∈{0,1,±2,5} and different FMM displacements Δx=Δy∈{0,0.15,0.25} mm. In addition, all the diffraction patterns are obtained with stochastic disturbances of the other two parameters: (i) beam waist size ω∈[0.45,0.55] mm; (ii) initial phase of OAM states φ0∈[0,2π]. Insets in (e1)–(e3) show the detailed profiles of the recorded intensity patterns.
    Fig. 3. Examples of the experimental diffraction intensity patterns for LG eigenstates with topological charge {0,1,±2,5} and different FMM displacements Δx=Δy{0,0.15,0.25}  mm. In addition, all the diffraction patterns are obtained with stochastic disturbances of the other two parameters: (i) beam waist size ω[0.45,0.55]  mm; (ii) initial phase of OAM states φ0[0,2π]. Insets in (e1)–(e3) show the detailed profiles of the recorded intensity patterns.
    Confusion matrix for the recognition of misaligned LG eigenstates ℓ∈{−5,−4,⋯,5} and p=0, with the curves of accuracy and loss as functions of epochs on the top right.
    Fig. 4. Confusion matrix for the recognition of misaligned LG eigenstates {5,4,,5} and p=0, with the curves of accuracy and loss as functions of epochs on the top right.
    Schematic diagram of a Bloch sphere constructed with |ℓ=±1⟩ bases. (a) An arbitrary state |ψ⟩ on the sphere and two adjacent superposition states ρA and ρB; (b) sphere is divided into 80,000 points corresponding to 80,000 states; the red box schematically indicates the position distribution of the nine selected superposition states for the experiments.
    Fig. 5. Schematic diagram of a Bloch sphere constructed with |=±1 bases. (a) An arbitrary state |ψ on the sphere and two adjacent superposition states ρA and ρB; (b) sphere is divided into 80,000 points corresponding to 80,000 states; the red box schematically indicates the position distribution of the nine selected superposition states for the experiments.
    Experimental results of hyperfine LG superposition states. (a)–(c) Examples of the recorded diffraction intensity patterns for OAM superposition states under different misaligned configurations. The collection of each diffraction pattern in the figure is carried out with stochastic disturbances of the other two parameters: (i) beam waist size ω∈[0.45,0.55] mm; (ii) initial phase of OAM states φ0∈[0,2π]. To show the image clearly, we reduce the contrast of the image by setting the values of I=0.3×max{I}, where I>0.3×max{I}; (d) confusion matrix for superposition states from Mode 1 to Mode 9 with the curves of accuracy and loss as functions of epochs on the top right.
    Fig. 6. Experimental results of hyperfine LG superposition states. (a)–(c) Examples of the recorded diffraction intensity patterns for OAM superposition states under different misaligned configurations. The collection of each diffraction pattern in the figure is carried out with stochastic disturbances of the other two parameters: (i) beam waist size ω[0.45,0.55]  mm; (ii) initial phase of OAM states φ0[0,2π]. To show the image clearly, we reduce the contrast of the image by setting the values of I=0.3×max{I}, where I>0.3×max{I}; (d) confusion matrix for superposition states from Mode 1 to Mode 9 with the curves of accuracy and loss as functions of epochs on the top right.
    Confusion matrix of LG modes with p=1. (a) Confusion matrix for LG eigenstates of ℓ∈{−2,−1,⋯,2} and p=1; (b) confusion matrix for superposition states from Mode 4 to Mode 8. Mode 4 to Mode 8 here represent the superposition states with p=1 at the same positions on the Bloch sphere in the case of p=0.
    Fig. 7. Confusion matrix of LG modes with p=1. (a) Confusion matrix for LG eigenstates of {2,1,,2} and p=1; (b) confusion matrix for superposition states from Mode 4 to Mode 8. Mode 4 to Mode 8 here represent the superposition states with p=1 at the same positions on the Bloch sphere in the case of p=0.
    ϕ=0.020πϕ=0.025πϕ=0.030π
    θ=0.520πMode 1Mode 4Mode 7
    θ=0.525πMode 2Mode 5Mode 8
    θ=0.530πMode 3Mode 6Mode 9
    Table 1. Nine Superposition States Combined by θ and ϕ
    Xiao Wang, Yufeng Qian, JingJing Zhang, Guangdong Ma, Shupeng Zhao, RuiFeng Liu, Hongrong Li, Pei Zhang, Hong Gao, Feng Huang, Fuli Li. Learning to recognize misaligned hyperfine orbital angular momentum modes[J]. Photonics Research, 2021, 9(4): B81
    Download Citation