基于频域约束独立成分分析的经验模态分解去噪方法
发布时间：2019-02-17 11:25
【摘要】：噪声污染是煤岩动力灾害电磁监测应用中需要解决的重要问题,去噪效果的好坏直接影响灾害预测的准确性。经验模态分解(Empirical Mode Decomposition,EMD)是目前电磁信号去噪中应用最多的一种方法,但当信号与噪声时频特征相近时,该算法存在严重的内蕴模态函数(Intrinsic Mode Function,IMF)混叠现象(即部分模态函数仍为信号与噪声的组合)。针对该问题,提出一种基于经验模态分解和频域约束独立成分分析的去噪方法,首先利用EMD将电磁信号分解为多个IMF分量,通过计算各分量与原信号间的互相关系数判断存在模态混叠现象过渡IMF,再以过渡IMF后续分量的频域为约束条件,对过渡IMF进行独立成分分析,去除过渡分量中的噪声;最后将去噪后的过渡分量与其后续分量进行重构,得到去噪后的信号。分别以含噪Ricker子波和现场电磁信号为例,利用信噪比定量验证了上述方法对处理现场电磁信号模态混叠问题的有效性,同时频域约束条件下的独立成分分析去噪收敛快、效率高,适合海量实时监测信号快速去噪使用。
[Abstract]:Noise pollution is an important problem to be solved in the application of electromagnetic monitoring of coal and rock dynamic disasters. The effect of noise removal directly affects the accuracy of disaster prediction. Empirical mode decomposition (Empirical Mode Decomposition,EMD) is one of the most widely used methods in electromagnetic signal denoising at present. However, when the signal and noise time and frequency characteristics are close, the algorithm has serious intrinsic mode function (Intrinsic Mode Function,. IMF) aliasing (i.e. partial mode function is still a combination of signal and noise). To solve this problem, a denoising method based on empirical mode decomposition (EMD) and frequency domain constrained independent component analysis (ICA) is proposed. Firstly, the electromagnetic signal is decomposed into multiple IMF components by EMD. By calculating the correlation number between each component and the original signal, the transition IMF, is judged to exist mode aliasing phenomenon. Then, the transition IMF is analyzed by independent component analysis (ICA) to remove the noise in the transition component by taking the frequency domain of the subsequent component of the transition IMF as the constraint condition. Finally, the de-noised transition component and its subsequent components are reconstructed to obtain the de-noised signal. Taking the noisy Ricker wavelet and the field electromagnetic signal as examples, the effectiveness of the above method in dealing with the field electromagnetic signal mode aliasing problem is quantitatively verified by using SNR, and the independent component analysis (ICA) denoising convergence is fast under the constraint of frequency domain. High efficiency and suitable for fast denoising of massive real-time monitoring signal.
【作者单位】： 福州大学环境与资源学院;
【基金】：国家自然科学基金资助项目(51604083)
【分类号】：TD326
[Abstract]:Noise pollution is an important problem to be solved in the application of electromagnetic monitoring of coal and rock dynamic disasters. The effect of noise removal directly affects the accuracy of disaster prediction. Empirical mode decomposition (Empirical Mode Decomposition,EMD) is one of the most widely used methods in electromagnetic signal denoising at present. However, when the signal and noise time and frequency characteristics are close, the algorithm has serious intrinsic mode function (Intrinsic Mode Function,. IMF) aliasing (i.e. partial mode function is still a combination of signal and noise). To solve this problem, a denoising method based on empirical mode decomposition (EMD) and frequency domain constrained independent component analysis (ICA) is proposed. Firstly, the electromagnetic signal is decomposed into multiple IMF components by EMD. By calculating the correlation number between each component and the original signal, the transition IMF, is judged to exist mode aliasing phenomenon. Then, the transition IMF is analyzed by independent component analysis (ICA) to remove the noise in the transition component by taking the frequency domain of the subsequent component of the transition IMF as the constraint condition. Finally, the de-noised transition component and its subsequent components are reconstructed to obtain the de-noised signal. Taking the noisy Ricker wavelet and the field electromagnetic signal as examples, the effectiveness of the above method in dealing with the field electromagnetic signal mode aliasing problem is quantitatively verified by using SNR, and the independent component analysis (ICA) denoising convergence is fast under the constraint of frequency domain. High efficiency and suitable for fast denoising of massive real-time monitoring signal.
【作者单位】： 福州大学环境与资源学院;
【基金】：国家自然科学基金资助项目(51604083)
【分类号】：TD326
【相似文献】
相关期刊论文 前9条
1 王建华;郑震;张延p，
本文编号：2425114
本文链接：https://www.wllwen.com/kejilunwen/kuangye/2425114.html
