窦性节律的QRS波终点处的晚电位已被用于诊断室性心动过速时的心肌缺血程度。用FFT对心电图进行频谱分析已用于检测这些微伏级的晚电位变异,但由于在数据较短时频率分辨率较低并产生谱泄漏,FFT频谱分析方法的应用受到一定的限制。本文提出一种基于自回归模型的最大熵谱(MEM)频率分析法。信号取自有室性心动过速症状的患者和无室速的心脏病患者及健康人,记录体表三导双极正交心电图,AD变换电路精度12位,采样率1000Hz,原始信号经低噪声高增益放大后送入AD Late potentials at the QRS end point of sinus rhythm have been used to diagnose myocardial ischemia during ventricular tachycardia. The FFT analysis of ECGs has been used to detect these micro-volts late potential variations, but the application of FFT spectral analysis methods is limited due to the lower frequency resolution and spectrum leakage at shorter data intervals. This paper presents a maximum entropy spectrum (MEM) frequency analysis method based on autoregressive model. The signal was taken from patients with symptoms of ventricular tachycardia and patients without heartbeat and with healthy heart rate. The three-lead bipolar orthogonal ECG was recorded. The accuracy of AD conversion circuit was 12 and the sampling rate was 1000Hz. The original signal was low Gain high noise amplification into AD