一维时序信号非局部均值滤波Python代码

以下是一维时序信号非局部均值滤波的Python代码：

import numpy as np
import scipy.stats as st

def nl_mean_filter(signal, window_size, h):
    """
    非局部均值滤波器
    :param signal: 一维时序信号
    :param window_size: 窗口大小
    :param h: 平滑参数
    :return: 平滑后的信号
    """
    filtered_signal = np.zeros_like(signal)

    for i in range(len(signal)):
        weights = []
        values = []

        # 遍历所有窗口
        for j in range(len(signal)):
            if j == i:
                continue

            # 计算窗口内的距离
            distance = np.sum((signal[j:j+window_size] - signal[i:i+window_size])**2)

            # 计算权重
            weight = st.norm.pdf(distance, scale=h)

            # 加入窗口值和权重
            values.append(signal[j])
            weights.append(weight)

        # 计算平滑值
        weighted_avg = np.average(values, weights=weights)

        # 将平滑值加入平滑信号
        filtered_signal[i] = weighted_avg

    return filtered_signal

使用方法如下：

import matplotlib.pyplot as plt

# 生成一维时序信号
signal = np.random.randn(100)

# 非局部均值滤波器平滑
window_size = 10
h = 0.5
filtered_signal = nl_mean_filter(signal, window_size, h)

# 绘制信号和平滑后的信号
plt.plot(signal, label='Original Signal')
plt.plot(filtered_signal, label='Filtered Signal')
plt.legend()
plt.show()

输出结果如下：