从临床采集的EMGdi信号一般非常微弱,而且会受到心电信号的干扰。本文使用能量比例阈值方法去除ECG干扰系数。首先,对受干扰的EMGdi信号进行小波变换,并且把小波系数转换为能量时谱;其次,能量比例阈值算法针对心电系数高于周围两边信号的特点,使用比例阈值对心电信号进行滤除;最后,小波能量时谱系数转换为小波系数,对小波系数进行重构得到降噪后的膈肌信号。对模拟仿真EMGdi信号与临床采集的EMGdi信号进行降噪,其结果表明本文的方法有效地降低了心电干扰并且保留EMGdi信号的特征。 EMGdi signals collected from the clinic are generally very weak, and may be disturbed by ECG signals. This article uses an energy proportional threshold method to remove ECG interference coefficients. Firstly, the disturbed EMGdi signals are transformed by wavelet and the wavelet coefficients are transformed into the energy spectrum. Secondly, the energy proportion threshold algorithm is used to filter the ECG signals by using the proportional threshold for the feature that the ECG coefficient is higher than the surrounding signals. Finally, the time-frequency coefficients of wavelet energy are converted into wavelet coefficients and the wavelet coefficients are reconstructed to obtain the denoised diaphragmatic signals. The simulated EMGdi signal and the clinically collected EMGdi signal are denoised. The results show that the proposed method effectively reduces ECG interference and preserves the features of EMGdi signal.