Empirical mode decomposition based background removal and de-noising in polarization interference imaging...

Area of Science:

• Optical Engineering
• Signal Processing
• Data Analysis

Background:

• Polarization Interference Imaging Interferometer (PIIS) generates data requiring background removal and de-noising.
• Noise suppression is crucial for accurate spectral analysis in PIIS data.

Purpose of the Study:

• To implement and validate Empirical Mode Decomposition (EMD) for background removal and de-noising PIIS data.
• To compare EMD-based methods with wavelet and windowed Fourier transform techniques.
• To determine optimal parameters for EMD thresholding de-noising.

Main Methods:

• Empirical Mode Decomposition (EMD) for adaptive background removal and de-noising.
• Numerical simulations to validate the effectiveness of EMD methods.
• Comparative analysis with wavelet and windowed Fourier transform de-noising methods.
• Evaluation using Signal-to-Noise Ratio (SNR), spectral resolution, and peak values.

Main Results:

• EMD-based methods effectively remove background and noise from PIIS data.
• The assumption that noise resides in the first intrinsic mode function (IMF) was verified.
• EMD demonstrated higher de-noising efficiency for Gaussian noise compared to Poisson noise.
• Wavelet and windowed Fourier transform methods were also analyzed for comparison.

Conclusions:

• Adaptive and robust EMD-based methods are highly effective for background removal and de-noising in PIIS.
• The developed EMD approach successfully yields background-free interferograms and noise-free spectra.
• EMD offers a reliable solution for enhancing the quality of PIIS data.