Spectroscopy and Spectral Analysis, Volume. 45, Issue 2, 344(2025)
Fast and Adaptive Raman Spectroscopy Baseline Correction Algorithm Based on the Principle of the Minus-Weighted Iterative Adjustment Least Square Method (MWIALS)
Raman spectroscopy is a non-destructive spectral analysis technique that obtains molecular structure information of substances by analyzing the frequency changes of scattered light. Baseline correction is a key step in enhancing spectral data quality, as it removes background signals and unrelated noise to highlight and purify the target signal. Traditional Raman spectroscopy applications do not require high timeliness for baseline correction. Still, in recent years, applications such as flow Raman and endoscopic Raman, which require real-time processing of spectral data, have increased, placing higher demands on the speed and accuracy of baseline correction. Traditional methods, such as iterative polynomial fitting and wavelet transform, have time, accuracy, or adaptability deficiencies. This study developed a fast adaptive baseline correction algorithm based on the Minus-Weighted Iterative Adjustment Least Square Method (MWIALS). The main principle is to extract the set of negative values and assign them higher weights, continuously adjust the baseline during the iteration process, and set parameter thresholds to exit the loop to achieve fast and accurate baseline correction. We also proposed two parameter selection strategies: Fixed Parameter (FMWIALS), suitable for rapid processing of batch homogeneous spectra, and Adaptive Parameter (AMWIALS), suitable for adaptive processing of heterogeneous spectra. The algorithm was applied to flow Raman spectral analysis of particulate matter, and the results showed that compared to other mainstream algorithms, it was significantly more efficient in practical spectral processing (average time of 47 milliseconds per spectrum) and achieved higher accuracy and adaptability. This algorithm can meet the real-time spectral processing needs in biological sample detection for flow Raman and endoscopic Raman applications, providing strong support for the further application of Raman spectroscopy technology.
Get Citation
Copy Citation Text
XU Jia-yang, MENG Si-yu, ZHANG Zhi-wei, CHEN Hong-yi, MA Yu-ting, WANG Ce, QI Xiang-dong, HU Hui-jie, SONG Yi-zhi. Fast and Adaptive Raman Spectroscopy Baseline Correction Algorithm Based on the Principle of the Minus-Weighted Iterative Adjustment Least Square Method (MWIALS)[J]. Spectroscopy and Spectral Analysis, 2025, 45(2): 344
Received: Jan. 16, 2024
Accepted: May. 21, 2025
Published Online: May. 21, 2025
The Author Email: HU Hui-jie (huhj@sibet.ac.cn)