APPLICATION OF FOURIER TRANSFORMS TO ANALYSIS OF SPECTRAL DATA

A computer method for automatically identifying the peaks in complex spectra and determining their centers and areas is described. The principal feature of the method is a data smoothing technique employing Fourier transforms. The smoothing eliminates most of the random fluctuations without effecting the spectral resolution and makes identification of maxima using a zero slope criterion possible. Using the same Fourier transform with different constants it is possible with a second transformation to improve the spectral resolution. The computer program has been written in FORTRAN IV for the MIT IBM 360 model 65 computer and also for the Toshiba Electric Co. GE 635 computer. The complete analysis of a 4096 channel spectrum containing several hundred peaks requires about 75 sec of computation time. © 1969.