An Automated Baseline Correction Method for Raman Spectra Based on Piecewise Polynomial Fitting With Adaptive Window

Baseline background tends to cause the Raman spectral signal to be hidden or distorted, making it difficult to accurately acquire spectral information. Although polynomial fitting has been widely proven to be an effective baseline correction method, it is difficult to achieve both high speed and high accuracy. In this paper, we propose a piecewise polynomial fitting method based on a sliding adaptive window(S-ModPoly). S-ModPoly relies on the adaptive width selection of a sliding window to automatically split the original spectrum into many different length segments containing complete peak information. Low-order iterative polynomial fitting is performed for each segment separately, which greatly reduces the computational effort while improving the baseline fitting accuracy. By further correcting the piecewise fitting results, the discontinuities between different intervals after piecewise fitting are eliminated. Furthermore, the true intensity information of the spectral signal is retained. Here, S-ModPoly is compared with three representative automated methods. The experimental results show that the S-ModPoly method has lower mean error and higher stability in a very short time (less than 20 ms). Additionally, the experimental results on measured Raman spectra demonstrate the effectiveness of the method in automatically processing various real spectral baselines. It performs well in both low and high intensity background baselines.