THE APPLICATION IN PROCESSING ANALYTICAL-CHEMISTRY SIGNALS OF A CARDINAL SPLINE APPROACH TO WAVELETS

The mth order B-spline N-m(x) with integer knots generates a multireso lution analysis, ...subset of V(-1)subset of V-0 subset of ..., with t he mth order of approximation in wavelet transformation (WT). We have compared the WT method with Fourier transformation processing analytic al chemistry signals in detail and have found that the WT method has m any advantages. This method can directly provide the frequency domain distribution of a signal in a time domain. The algorithms are simpler and take less time in operation, and one needs only to decompose the d igital signal rather than to undertake many transformations. Moreover, it is not necessary to preprocess the original signal to analyze the pattern of data and to know the statistical character of the noise. Wh en the signal-to-noise ratio is 0.2, the processed results of theoreti cal and experimental data can still be satisfactory.