一种MCG信号的自适应预测滤波方法

Adaptive Prediction Filter for MCG Signal

与心电图相比较,心磁检测具有无需电极,对某些局部心肌电流高度敏感,可以用于早期心脏疾病的诊断。但心脏磁场信号在检测过程中会被噪声所污染,使得信号本身的可辨识性降低,因此需要对该信号进行降噪处理。在单通道信号采集系统无噪声参考输入端的情况下采用自适应滤波方法,需要对待处理信号进行线性预测,本文提出的改进LMS(Least Mean Square)算法的自适应预测滤波器,无需噪声参考信号即可对心磁信号进行滤波,通过三种不同噪声的滤波仿真结果可见,采用自适应预测滤波器处理后明显提高了信噪比,具有一定的学术意义和实用价值。

Compared with Electrocardiograph, the Magnetocardiography （MCG） signal dtetection Doesn’t need electrodes, is highly sensetive with part cardiac muscle currents, and can be used in the diagnosis of early cardiopathy. But Magnetocardiography signals would be polluted by the noise or disturbance during the detection, which leads to lower in character, quality of identification. Without noise reference input, it need linear prediction in the single detection channel process. Hence magnetocardiography signals required series of noise reduction for further investigation. It has developed two new adaptive prediction filters based on LMS algorithm and improved LMS algorithm, and both algorithms didn＇ t need noise signal reference input for filtering. The simulated filtering results of three different noise signals demonstrate that it improves the SNR （ Signal Noise Ratio） of magnetocardiography signals significantly through these methods.