Calibration transfer between two near-infrared spectrometers based on a wavelet packet transform

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• Near-infrared (NIR) spectroscopy is widely used for chemical analysis.
• Spectral variability between instruments necessitates standardization for accurate data transfer.
• Existing methods for calibration transfer can be complex and computationally intensive.

Purpose of the Study:

• To propose a novel standardization algorithm, WPTS, for spectral transfer between NIR spectrometers.
• To evaluate the performance of WPTS against existing calibration transfer methods.
• To demonstrate a simpler and computationally efficient approach to spectral standardization.

Main Methods:

• Development of the Wavelet Packet Transform Standardization (WPTS) algorithm.
• Utilizing entropy criteria and wavelet packet transform (WPT) for optimal WPT-tree generation.
• Establishing a transfer matrix based on selected WPT-tree nodes.
• Validation using spectra from standardization samples measured on two different NIR spectrometers.

Main Results:

• WPTS successfully transfers spectra between NIR spectrometers, enabling data to be treated as if measured on the target instrument.
• The proposed WPTS achieves transfer performance comparable to the best existing method (WTS1).
• WPTS outperforms WTS2 and classical PDS methods in transfer accuracy.
• WPTS employs a simpler variable selection technique and exhibits lower computational cost.

Conclusions:

• WPTS provides an effective and efficient method for spectral standardization in NIR spectroscopy.
• The algorithm simplifies the process of calibration transfer while maintaining high accuracy.
• WPTS offers a valuable alternative for researchers and practitioners dealing with multi-instrument NIR data.