Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 62, Issue 9, 0930003(2025)

High-Resolution Computational Spectrometer Based on Automatic Differentiation Optimization Algorithm

Yangfan Qi1... Junrui Liang1, Jun Li1, Jiangming Xu1, Jinyong Leng1,2,3 and Pu Zhou1,* |Show fewer author(s)
Author Affiliations
  • 1College of Advanced Interdisciplinary Studies, National University of Defense Technology, Changsha 410073, Hunan , China
  • 2Nanhu Laser Laboratory, National University of Defense Technology, Changsha 410073, Hunan , China
  • 3Hunan Provincial Key Laboratory of High Energy Laser Technology, National University of Defense Technology, Changsha 410073, Hunan , China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (9)
    Equations (5)
    References (53)
    Tools
    Paper Information
    Topics
    Figures & Tables(9)
    The speckle patterns at the output end of the multimode fiber. (a) 1550.000 nm; (b) 1550.001 nm; (c) 1550.002 nm; (d) 1550.003 nm.

    Fig. 1. The speckle patterns at the output end of the multimode fiber. (a) 1550.000 nm; (b) 1550.001 nm; (c) 1550.002 nm; (d) 1550.003 nm.

    Download full size
    Experimental setup diagram

    Fig. 2. Experimental setup diagram

    Download full size
    Correlation of speckle patterns with wavelength variation and reconstructed spectra. (a) Correlation of speckle patterns with wavelength variation; (b) AD; (c) SA; (d) CS-L1; (e) CS-L2; (f) NN

    Fig. 3. Correlation of speckle patterns with wavelength variation and reconstructed spectra. (a) Correlation of speckle patterns with wavelength variation; (b) AD; (c) SA; (d) CS-L1; (e) CS-L2; (f) NN

    Download full size
    The reconstruction results of spectra with a relative sparsity ratio of 3% using different algorithms. (a) AD; (b) SA; (c) CS-L1; (d) CS-L2; (e) NN

    Fig. 4. The reconstruction results of spectra with a relative sparsity ratio of 3% using different algorithms. (a) AD; (b) SA; (c) CS-L1; (d) CS-L2; (e) NN

    Download full size
    The relative error and time consumption of five algorithms in reconstructing different Lorentzian bandwidth spectra. (a) Reconstruction error for spectral optimization by five algorithms; (b) time for spectral optimization by five algorithms

    Fig. 5. The relative error and time consumption of five algorithms in reconstructing different Lorentzian bandwidth spectra. (a) Reconstruction error for spectral optimization by five algorithms; (b) time for spectral optimization by five algorithms

    Download full size
    The reconstruction error and time of four types of spectra using different algorithms. (a) Reconstruction error; (b) time

    Fig. 6. The reconstruction error and time of four types of spectra using different algorithms. (a) Reconstruction error; (b) time

    Download full size
    The results of the AD-optimized spectra. (a) The tunable narrow peak in the working range; (b) the continuous broad spectrum; (c) the broad spectrum with a peak-like narrow spectrum; (d) the broad spectrum with an embedded peak-like narrow spectrum

    Fig. 7. The results of the AD-optimized spectra. (a) The tunable narrow peak in the working range; (b) the continuous broad spectrum; (c) the broad spectrum with a peak-like narrow spectrum; (d) the broad spectrum with an embedded peak-like narrow spectrum

    Download full size

    • Table 1. The reconstruction error and time of different Lorentz bandwidth spectra using five algorithms

      View table

      Table 1. The reconstruction error and time of different Lorentz bandwidth spectra using five algorithms

      AlgorithmIterativeNon-iterative
      ADSACS-L1CS-L2NN
      ErrorMean value0.10560.11711.15570.07670.4851
      Standard deviation1.7×10-31.9×10-30.56283.3×10-32.7×10-2
      Time /sMean value1.684108.6202.5101.4

      Train： 27 h

      Predict： 0.0194 s

      Standard deviation0.555129.8649.7117.600.0131

    • Table 2. The reconstruction error and time of different in four types of spectra using five algorithms

      View table

      Table 2. The reconstruction error and time of different in four types of spectra using five algorithms

      AlgorithmIterativeNon-iterative
      ADSACS-L1CS-L2NN
      ErrorMean value0.12750.13871.3410.22360.6507
      Standard deviation3.8×10-31.3×10-20.21317.7×10-30.1772
      Time /sMean value1.64870.19180.254.35

      Train： 27 h

      Predict： 0.0117 s

      Standard deviation0.449523.9339.432.7150.0073
    Tools

    Get Citation

    Copy Citation Text

    Yangfan Qi, Junrui Liang, Jun Li, Jiangming Xu, Jinyong Leng, Pu Zhou. High-Resolution Computational Spectrometer Based on Automatic Differentiation Optimization Algorithm[J]. Laser & Optoelectronics Progress, 2025, 62(9): 0930003

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Set citation alerts for article
    Save article for my favorites
    Paper Information

    Category: Spectroscopy

    Received: Aug. 19, 2024

    Accepted: Sep. 23, 2024

    Published Online: Apr. 23, 2025

    The Author Email: Pu Zhou (zhoupu203@163.com)

    DOI:10.3788/LOP241858

    CSTR:32186.14.LOP241858

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology