spectral entropy python代码

以下是计算信号的谱熵的 Python 代码示例：

import numpy as np
from scipy import signal

def spectral_entropy(x, sf, method='fft', nperseg=None, normalize=False):
    """Calculate spectral entropy of a signal.

    Args:
        x (np.ndarray): Input signal.
        sf (float): Sampling frequency.
        method (str): Method to use for spectral density estimation.
            Options are 'fft' for FFT and 'welch' for Welch periodogram.
            Default is 'fft'.
        nperseg (int): Length of each segment in Welch method. Only used if
            method is 'welch'. If None, it is set to half the length of x.
            Default is None.
        normalize (bool): Whether to normalize the spectral density to sum to
            1. Default is False.

    Returns:
        float: Spectral entropy of the signal.

    """
    if method == 'fft':
        f, pxx = signal.periodogram(x, sf)
    elif method == 'welch':
        if nperseg is None:
            nperseg = len(x) // 2
        f, pxx = signal.welch(x, sf, nperseg=nperseg)
    else:
        raise ValueError('Invalid method specified.')

    if normalize:
        pxx /= np.sum(pxx)

    return -np.sum(pxx * np.log2(pxx))

该函数使用 `scipy` 库中的 `signal` 模块计算信号的谱熵。可以选择使用 FFT 或 Welch 周期图法进行谱密度估计。如果选择使用 Welch 方法，则可以指定每个段的长度。如果将 `normalize` 参数设置为 `True`，则会将谱密度归一化为总和为 1。

使用示例：

x = np.random.randn(1000)  # Generate random signal
sf = 1000  # Sampling frequency
se = spectral_entropy(x, sf)  # Compute spectral entropy
print(se)