Vol. 37, issue 11, article # 1

Shcherbakov A. P. Neural network for fitting vibrational-rotational line contours in high-resolution spectra. // Optika Atmosfery i Okeana. 2024. V. 37. No. 11. P. 897–904. DOI: 10.15372/AOO20241101 [in Russian].
Copy the reference to clipboard

Abstract:

The article is devoted to the problem of automation of fitting lines parameters in high-resolution spectra recorded at modern spectrometers. When fitting a model contour, due to the presence of many local minima in the minimized standard deviation, a sufficiently accurate initial approximation of line profile parameters is required. The article proposes a method for finding a sufficiently accurate initial approximation of line profile parameters for the convergence of the fitting process. The method is based on the Kohonen neural network. Tests and comparison of other algorithms and networks for solving this problem are carried out. The suggested method can be used to process vibrational-rotational spectra and obtain databases for solving problems of atmospheric optics, molecular physics, and engineering problems.

Keywords:

absorption line parameters, Fourier transform spectrometer, Voigt profile, automatic processing, neural network

Figures:

References:

1. Kuryak A.N., Tikhomirov B.A. Vliyanie vodyanogo para na pogloshchenie izlucheniya 266 nm alyuminievym opticheskim zerkalom // Optika atmosf. i okeana. 2018. V. 31, N 10. P. 791–793. DOI: 10.15372/AOO20181003.
2. Vasilenko I.A., Naumenko O.V. Ekspertnyi spisok linii pogloshcheniya molekuly 32S16O2 v diapazone 2000–3000 cm-1 // Optika atmosf. i okeana. 2020. V. 33, N 5. P. 342–346. DOI: 10.15372/AOO20200503; Vasilenko I.A., Naumenko O.V., Horneman V.-M. Expert list of absorption lines of the SO2 molecule in the 2000–3000 cm-1 spectral region // Atmos. Ocean. Opt. 2020. V. 33, N 5. P. 443–448.
3. Romashenko O.P., Kornev A.S., Zon B.A. Pogloshchenie lazernogo izlucheniya v atmosfere Titana // Optika atmosf. i okeana. 2020. V. 33, N 5. P. 329–333. DOI: 10.15372/AOO20200501; Romashenko O.P., Kornev A.S., Zon B.A. Laser radiation absorption in the atmosphere of Titan // Atmos. Ocean. Opt. 2020. V. 33, N 5. P. 439–442.
4. Shcherbakov A.P., Pshenichnicov A.M. Computer-aided system for automatic peak searching and contour fitting in molecular spectra // Proc. SPIE. 2000. N 4341. P. 60–63. DOI: 10.1117/12.411924.
5. Kruglova T.V., Shcherbakov A.P. Avtomaticheskii poisk linii v molekulyarnykh spektrakh na osnove metodov neparametricheskoi statistiki. Regulyarizatsiya v otsenke parametrov spektral'nykh linii // Opt. i spektroskop. 2011. V. 111, N 3. P. 383–386.
6. Tikhonov A.N., Arsenin V.Ya. Metody resheniya nekorrektnykh zadach. M.: Nauka, Glavnaya redaktsiya fiziko-matematicheskoi literatury, 1979. 285 p.
7. Levin L.L. Vvedenie v teoriyu raspoznavaniya obrazov: ucheb. posobie. Tomsk: TGU, 1982, 2004, 2008. 97 p.
8. Aizerman M.A., Braverman E.I., Rozonoer L.I. Metod potentsial'nykh funktsii v zadachakh obucheniya mashin. M.: Nauka, 1970. 384 p.
9. Khaikin S. Neironnye seti: polnyi kurs / per. s angl. M.: Vil'yams, 2008. 1103 p.
10. Fausett Laurene V. Fundamentals of Neural Networks: Architectures, Algorithms and Applications. Pearson Education, 2006. 467 p.