Bandwidth correction of spectral measurement based on Levenberg-Marquardt algorithm with improved Tikhonov regularization.

Spectroscopic data often suffers from spectral distortions due to the broadening effects of the spectrometer. In this paper, the issue of bandwidth correction is transformed into a multiparameter optimization problem, and an improved bandwidth correction method based on the Levenberg-Marquardt algorithm with improved Tikhonov regularization is presented. Example LED, Raman, and compact fluorescent lamp spectra were corrected with the proposed method, the Richardson-Lucy method, and the maximum a posterior method, and the correction errors and runtimes under the same stopping conditions of each method were compared. The experimental results show that the proposed method can be applied to different spectra with excellent correction.