基于小波阈值算法的中心静脉压信号去噪研究

Study on Denoising of Central Venous Pressure Signal Based on Wavelet Threshold Algorithm

为去除无创中心静脉压监测仪采集的中心静脉压信号中的噪声,提出了一种基于小波阈值的去噪算法。针对采集过程中出现的基线漂移和高频噪声,通过选取合适的小波基、小波分解层数和阈值方法,对中心静脉压信号进行小波阈值处理,并与经过傅里叶变换和维纳滤波去噪后的结果进行对比。在信噪比、均方根差等评价指标上,小波阈值的效果更好。选择sym3小波基对中心静脉压信号进行6层小波分解,设置硬阈值所得到的效果更好。该算法能够有效去除采集过程中产生的噪声信号,信噪比达到90.1942dB,均方根差接近0,与其他算法相比去噪效果更好。

In order to remove the noise in the central venous pressure signal collected by the non-invasive central venous pressure monitor,a wavelet threshold based denoising algorithm is proposed.Aiming at the baseline drift and high-lrequency noise during the acquisition process,the wavelet threshold treatment of the central venous pressure signal was performed by selecting the appropriate wavelet base,wavelet decomposition layer and threshold method,and compared with the results after denoising by Fourier transform and Wiener filter.In terms of signal-to-noise ratio and rms difference,the wavelet threshold has a better effect.Selecting the Sym3 wavelet base for 6-layer wavelet decomposition of the central venous pressure signal,and setting the hard threshold is better.The algorithm can effectively remove the noise signal generated during the acquisition process,the signal-to-noise ratio reaches 90.1942 dB,and the rms difference is close to 0,which has a better denoising effect than other algorithms.