Baseline correction for Raman spectra using an improved asymmetric least squares method

Baseline shifts exist in many types of Raman spectrometers. Acquired spectra normally contain the desired signals as well as undesirable elements such as background noise. In this paper, an improved asymmetric least squares (IAsLS) method has been proposed for the baseline correction of Raman spectra. The baseline correction algorithm is initiated by the raw spectrum baseline, and this baseline can be estimated using a polynomial fitting method. For the simulated Raman spectra, the performance of the proposed algorithm was evaluated and compared with the asymmetric least squares (AsLS) method and Jiang's method. The results showed that it is improved by sixteen fold and nine fold respectively. This proposed IAsLS method was successfully applied to practical Raman spectral data and the results in the paper indicate that the baseline of Raman spectra can be automatically subtracted.