We propose a numerical method to eliminate the background noise from a line spectrum. The background noise is discriminated from signal components by applying a singular value decomposition (SVD) method to a cyclic matrix that is constructed from the spectrum data. The first column of the cyclic matrix is given by the Fourier transform of the spectrum data. Unlike conventional smoothing procedures, the new method gives no distortion to the signal waveform. The details of the method are described. Experimental results applied to FT-Raman spectrum data of carbon tetrachloride are also shown.
Tetsuo Iwata and Jun Koshoubu, "New Method to Eliminate the Background Noise from a Line Spectrum," Appl. Spectrosc. 48, 1453-1456 (1994)