Curve fitting using wavelet transform for resolving simulated overlapped spectra

Curve fitting must be considerably facilitated if reliable and accurate initial estimates of the number of peaks, individual peak positions, areas, and widths are known at the outset. The most important values for input to curve fitting route are the number of peaks and their positions. One of the main drawbacks involved is that as the bands become more overlapped, or the number of overlapped bands increases, the problem becomes progressively more illconditioned. As a consequence, small errors in the data, or errors in the estimates can be magnified, ultimately resulting in large errors in the final model. In addition, very high noise level of the analytical signal also has significant effect on the fitted results. In this work, curve fitting using wavelet transform for peak finding and an enhancement of signal-to-noise ratio was proposed, in which wavelet transform was performed prior to curve fitting to enhance noise level for subsequent fitting and to determine the number of peaks and corresponding parameters. Accordingly, the fitted conditions can be improved to the point that very accurate results could be acquired even for the simulated overlapped bands with higher noise level.