改进的自适应噪声总体集合经验模态分解在光谱信号去噪中的应用

发布时间：2018-06-12 05:24

本文选题：无创血糖检测 + 近红外光谱　；参考：《光学精密工程》2016年07期

【摘要】：针对近红外无创血糖检测过程中噪声对血糖浓度模型精度和稳定性的影响,提出用自适应噪声总体集合经验模态分解方法实现近红外光谱信号的去噪;同时,根据原始信号曲率和分解后本征模态函数(IMFs)曲率间的离散弗雷歇距离选择相关模态。首先,将自适应噪声的总体集合经验模态分解方法引入近红外光谱去噪过程,介绍了经验模态分解、集合经验模态分解、互补集合经验模态分解及自适应噪声总体集合经验模态分解的基本原理及具体实现过程。然后,应用基于曲率和离散弗雷歇距离的自适应噪声总体集合经验模态分解改进算法对仿真信号和光谱信号进行去噪,并将其标准差和信噪比作为评价指标。实验结果表明:应用提出的方法得到的血糖浓度近红外光谱数据其标准差为0.179 4,信噪比为19.117 5dB,实现了信号与噪声的分离,改善了重构信号质量,具有良好的自适应性,可以有效识别并提取有用信息。
[Abstract]:Aiming at the effect of noise on the accuracy and stability of blood glucose concentration model during NIR noninvasive blood glucose detection, an adaptive noise ensemble empirical mode decomposition method is proposed to de-noise the near-infrared spectral signal. According to the discrete Frechch distance between the curvature of the original signal and the decomposed eigenmode function IMFs, the correlation modes are selected. Firstly, the aggregate empirical mode decomposition method of adaptive noise is introduced into the near infrared spectrum denoising process, and the empirical mode decomposition (EMD) and set empirical mode decomposition (EMD) are introduced. The basic principle and realization process of complementary set empirical mode decomposition and adaptive noise aggregate empirical mode decomposition. Then, an improved adaptive noise set empirical mode decomposition algorithm based on curvature and discrete Frechch distance is applied to de-noising the simulated and spectral signals, and its standard deviation and signal-to-noise ratio (SNR) are taken as the evaluation indexes. The experimental results show that the proposed method has a standard deviation of 0.179 4 and a signal-to-noise ratio of 19.117 5 dB. The separation of signal and noise is realized, the quality of reconstructed signal is improved, and the self-adaptability of the reconstructed signal is improved. It can effectively identify and extract useful information.
【作者单位】：哈尔滨工业大学电气工程及自动化学院;
【基金】：哈尔滨市科技创新人才专项资金资助项目(No.2014RFXXJ065) 哈尔滨工业大学理工医交叉学科基础研究培育计划资助项目(No.HIT.IBRSEM.201307)
【分类号】：R587.1;O657.33;R446.11

【相似文献】