A UV-visible absorption spectrum denoising method based on EEMD and an improved universal threshold filter.

Area of Science:

• Spectroscopy
• Data Science
• Environmental Science

Background:

• Ultraviolet-visible (UV-Vis) spectroscopy is crucial for water quality analysis.
• Raw spectral data often suffers from interference, necessitating effective denoising.
• Inaccurate denoising hinders subsequent quantitative analysis and information extraction.

Purpose of the Study:

• To introduce a new denoising method for UV-Vis absorption spectra.
• To evaluate the proposed method's performance against existing techniques.
• To improve the accuracy and capability of spectral analysis in water quality monitoring.

Main Methods:

• Ensemble Empirical Mode Decomposition (EEMD) to decompose spectral signals into intrinsic mode functions (IMFs).
• Spearman's rank correlation coefficient to identify noise-dominated IMFs.
• Improved universal threshold filtering applied to noise-dominated IMFs for noise reduction.
• Reconstruction of the denoised UV-Vis absorption spectrum.

Main Results:

• The proposed EEMD-based denoising method significantly outperforms wavelet-based threshold denoising methods on synthetic signals.
• The method demonstrated robust performance in denoising measured UV-Vis absorption spectra with varying signal-to-noise ratios (SNR).
• Spearman's rho effectively distinguished between noise and signal-dominated IMFs.

Conclusions:

• The EEMD-based denoising method offers a superior approach for processing noisy UV-Vis spectral data.
• This technique enhances the reliability of spectral analysis for water quality parameters.
• The proposed method holds promise for advancing spectral analysis capabilities in environmental monitoring.