基于cv2实现的四种图片均值滤波器

基于cv2实现的四种图片均值滤波器

一、算数均值

import numpy as np
from cv2 import cv2

def a_mean(img, kernel_size):
    G_mean_img = np.zeros(img.shape)
    k = int((kernel_size - 1) / 2)
    for i in range(img.shape[0]):
        for j in range(img.shape[1]):
            if i < k or i > (img.shape[0] - k - 1) or j < k or j > (img.shape[1] - k - 1):
                G_mean_img[i][j] = img[i][j]
            else:
                for n in range(kernel_size):
                    for m in range(kernel_size):
                        G_mean_img[i][j] += float(1 / (kernel_size * kernel_size) * img[i - k + n][j - k + m])
    G_mean_img = np.uint8(G_mean_img)
    return G_mean_img

if __name__ == "__main__":
    img = cv2.imread('data.jpg', 0)
    G_mean_img_3 = a_mean(img, kernel_size=3)
    cv2.imshow("source", img)
    cv2.imshow("a", G_mean_img_3)
    cv2.waitKey(0)

    print("算数均值滤波完成")

二、几何均值滤波

import numpy as np
from cv2 import cv2

def b_mean(img, kernel_size):
    G_mean_img = np.ones(img.shape)
    k = int((kernel_size - 1) / 2)
    for i in range(img.shape[0]):
        for j in range(img.shape[1]):
            if i < k or i > (img.shape[0] - k - 1) or j < k or j > (img.shape[1] - k - 1):
                G_mean_img[i][j] = img[i][j]
            else:
                for n in range(kernel_size):
                    for m in range(kernel_size):
                        G_mean_img[i][j] *= float(img[i - k + n][j - k + m])
                G_mean_img[i][j] = pow(G_mean_img[i][j], 1 / (kernel_size * kernel_size))
    G_mean_img = np.uint8(G_mean_img)
    return G_mean_img

if __name__ == "__main__":
    img = cv2.imread('data.jpg', 0)
    G_mean_img_3 = b_mean(img, kernel_size=5)
    cv2.imshow("source", img)
    cv2.imshow("b", G_mean_img_3)
    cv2.waitKey(0)

    print("几何均值滤波完成")

三、调和均值滤波

import numpy as np
from cv2 import cv2

def c_mean(img, kernel_size):
    G_mean_img = np.zeros(img.shape)
    k = int((kernel_size - 1) / 2)
    for i in range(img.shape[0]):
        for j in range(img.shape[1]):
            if i < k or i > (img.shape[0] - k - 1) or j < k or j > (img.shape[1] - k - 1):
                G_mean_img[i][j] = img[i][j]
            else:
                for n in range(kernel_size):
                    for m in range(kernel_size):
                        if img[i - k + n][j - k + m] == 0:
                            G_mean_img[i][j] = 0
                            break
                        else:
                            G_mean_img[i][j] += 1 / float(img[i - k + n][j - k + m])
                    else:
                        continue
                    break
                if G_mean_img[i][j] != 0:
                    G_mean_img[i][j] = (kernel_size * kernel_size) / G_mean_img[i][j]
    G_mean_img = np.uint8(G_mean_img)
    return G_mean_img

if __name__ == "__main__":
    img = cv2.imread('data.jpg', 0)
    G_mean_img_3 = c_mean(img, kernel_size=5)
    cv2.imshow("source", img)
    cv2.imshow("c", G_mean_img_3)
    cv2.waitKey(0)

    print("调和均值滤波完成")

四、逆调和均值滤波

import numpy as np
from cv2 import cv2

def d_mean(img, kernel_size, Q):
    G_mean_img = np.zeros(img.shape)
    k = int((kernel_size - 1) / 2)

    for i in range(img.shape[0]):
        for j in range(img.shape[1]):
            if i < k or i > (img.shape[0] - k - 1) or j < k or j > (img.shape[1] - k - 1):
                G_mean_img[i][j] = img[i][j]
            else:
                result_top = 0
                result_down = 0
                for n in range(kernel_size):
                    for m in range(kernel_size):
                        if Q > 0:
                            result_top += pow(float(img[i - k + n][j - k + m]), Q + 1)
                            result_down += pow(float(img[i - k + n][j - k + m]), Q)
                        else:
                            if img[i - k + n][j - k + m] == 0:
                                G_mean_img[i][j] = 0
                                break
                            else:
                                result_top += pow(float(img[i - k + n][j - k + m]), Q + 1)
                                result_down += pow(float(img[i - k + n][j - k + m]), Q)
                    else:
                        continue
                    break
                else:
                    if result_down != 0:
                        G_mean_img[i][j] = result_top / result_down
    G_mean_img = np.uint8(G_mean_img)
    return G_mean_img

if __name__ == "__main__":
    img = cv2.imread('data.jpg', 0)
    G_mean_img_3 = d_mean(img, kernel_size=9, Q=-1.5)
    cv2.imshow("source", img)
    cv2.imshow("c", G_mean_img_3)
    cv2.waitKey(0)

    print("逆调和均值滤波完成")