光学学报, 2021, 41 (1): 0130001, 网络出版: 2020-07-31
生物医学检测中太赫兹光谱技术的算法研究 下载: 2172次特邀综述
Terahertz Spectroscopy Algorithms for Biomedical Detection
光谱学 太赫兹光谱技术 算法 信号降噪 数据重构 定性及定量分析 spectroscopy terahertz spectroscopy algorithms signal denosing data reconstruction qualitative and quantitative analysis
摘要
基于太赫兹波的非电离、非侵入性、高穿透性、高分辨率和光谱指纹特征,太赫兹光谱技术在生物医学领域具有巨大潜力。基于太赫兹光谱技术和不同的分析算法,不同研究小组实现了对混合物样品的定性、定量识别。然而,实际的生物混合物样品中通常包含水在内的不同成分,进而导致光谱的信噪比较差,导致最终的光谱分析结果误差较大。对于此类问题,降噪算法和重构算法是比较有效的解决办法。这些算法通过去除光谱数据中的无效信息或提取其中的有效信息来达到提高光谱信噪比的目的,最终结合分析算法实现对生物样本的高精度定性和定量识别。本文对近五年来应用于太赫兹光谱技术中的主要算法进行了归纳介绍,并总结了它们的优势和缺点。
Abstract
Based on the features of nonionization, noninvasiveness, high penetration, high resolution, and spectral fingerprinting of terahertz (THz) waves, terahertz spectroscopy has great potential in the biomedical field. Based on terahertz spectroscopy, combined with different analysis algorithms, different research groups have achieved qualitative and quantitative identification of mixture samples. However, actual biological mixture samples often comprise different components, including water, which results in poor spectral signal-to-noise ratio and large errors in the final spectral analysis results. For these problems, the use of noise reduction and reconstruction algorithms is effective solutions. These algorithms improve the signal-to-noise ratio of the spectrum by eliminating invalid information in the spectral data or extracting valid information. Finally, these algorithms can be combined with analysis algorithms to provide high-precision qualitative and quantitative identification of biological samples. In this paper, we discuss the main algorithms applied in terahertz spectroscopy over the past five years and summarize their advantages and disadvantages.
朱亦鸣, 施辰君, 吴旭, 彭滟. 生物医学检测中太赫兹光谱技术的算法研究[J]. 光学学报, 2021, 41(1): 0130001. Yiming Zhu, Chenjun Shi, Xu Wu, Yan Peng. Terahertz Spectroscopy Algorithms for Biomedical Detection[J]. Acta Optica Sinica, 2021, 41(1): 0130001.