Preprocessing optical spectra superposition for further decomposition

This paper presents research on the use of discrete Fourier transform algorithm to limit the amount of data without losing information. Such preprocessing of registered optical spectra is necessary in order to shorten the execution time of the complex spectral deconvolution algorithms. It was shown that the spectral reduction coefficient with the noise component is sensitive to the noise level and decreases asymptotically to the hardware signal oversampling ratio.