一维非局部均值滤波Python代码

以下是一维非局部均值滤波的Python代码：

import numpy as np
import cv2

def non_local_means_filter_1d(img, h, search_window, patch_size):
    filtered_img = np.zeros_like(img)
    img_padded = cv2.copyMakeBorder(img, patch_size//2, patch_size//2, 0, 0, cv2.BORDER_REPLICATE)
    for i in range(img.shape[0]):
        i_padded = i + patch_size//2
        weighted_sum = 0
        similarity_sum = 0
        for j in range(i-search_window, i+search_window+1):
            if j < 0 or j >= img.shape[0]:
                continue
            j_padded = j + patch_size//2
            patch1 = img_padded[i_padded-patch_size//2:i_padded+patch_size//2+1]
            patch2 = img_padded[j_padded-patch_size//2:j_padded+patch_size//2+1]
            similarity = np.exp(-np.sum((patch1-patch2)**2)/(h*h))
            weighted_sum += similarity * img[j]
            similarity_sum += similarity
        filtered_img[i] = weighted_sum / similarity_sum
    return filtered_img

其中，`img`是一维的输入图像，`h`是滤波参数，`search_window`是搜索窗口大小，`patch_size`是块大小。`filtered_img`是滤波后的图像。

函数首先对输入图像进行边缘填充，然后对每个像素点进行非局部均值滤波。对于每个像素点，搜索窗口内的像素点与当前像素点的块进行相似度计算，根据相似度加权求和得到滤波后的像素值。最后返回滤波后的图像。