Spectral Analysis Using the Wavelet Transform

Open Access
Publisher Website
Google Scholar
Abstract
We introduce a new signal processing technique to analyze noisy spectra. The method is based on the wavelet transform and employs the à trous algorithm. Noise determination and detection criteria are discussed in detail, together with pitfalls related to the use of wavelets in the analysis of spectra. Simulations are presented to demonstrate the power and the shortcomings of our method. We apply our technique to the case of continuum sources that show superposed interstellar or circumstellar absorption or emission bands that are shallow and broad. In particular, we analyze an L-band spectrum of the Herbig-Haro energy source HH 100 IRS. The analysis indicates the presence of a shallow emission band near 3.51 μm that is tentatively assigned to arise from aliphatic (CH2) vibrations.