writer：Xihui Bian*, Yajing Yan, Yue Hao, Mengxuan Ling, Weilu Tian*
keywords：Analytical instrument signals, denoising, smoothing, signal decomposition, signal modeling
source：期刊
specific source：Measurement, 2026, 268, 120567
Issue time：2026年
Background: Denoising has gained increasing attention with the miniaturization of analytical instruments and the microcosm of analytical objects. In order to enhance the signal-to-noise ratio of the signal, researchers have devoted considerable efforts to hardware modification and the development of denoising algorithms.
Results: Recent advances denoising methods for analytical instrument signals, including algorithms, strategies, evaluation, and applications have been summed. Denosing algorithms are categorized into four main groups: smoothing-based, decomposition-based, modeling-based, and deep learning. For each category, this review delves into the underlying principles and distinctive features, providing a thorough analysis of their respective strengths and weaknesses. In light of the shortcomings identified in current algorithms, this paper presents effective strategies to address these issues. Additionally, it introduces multiple indicators for evaluating the effectiveness of denoising and offers a comprehensive summary of the applications of denoising algorithms in different analytical instrument signals.Significance: The review will be very useful for denoising of analytical instrument signals and other signals that need denoising such as medical signals, seismic signals, acoustic signals, etc. It will systematic summary the principle, advantages, disadvantages and applications of existing denoising methods for analytical instrument signals, which is very helpful for the use and development of denoising methods. Chemometricians, materials chemists, medical scientists, and instrument builders can use denoising methods to solve data quality problems, improve analytical accuracy and reliability, and thus promote the development and application of related technologies.