利用Python和OpenCV分别计算原始图像的高斯金字塔，计算原始图像的高斯差分金字塔（DoG），计算原始图像的拉普拉斯金字塔（LoG），分别给出三个代码

好的，以下是三个代码实现：

1. 计算高斯金字塔

import cv2

# 读取原始图像
img = cv2.imread("image.png")

# 定义高斯金字塔层数和尺度
num_octaves = 4
num_scales = 5

# 计算高斯金字塔
gaussian_pyramid = []
k = pow(2, 1/num_scales)
for i in range(num_octaves):
    octave = []
    for j in range(num_scales):
        if i == 0 and j == 0:
            octave.append(img)
        elif j == 0:
            pre_img = gaussian_pyramid[-1][-3]
            cur_img = cv2.pyrDown(pre_img)
            octave.append(cur_img)
        else:
            pre_img = octave[-1]
            cur_img = cv2.GaussianBlur(pre_img, (0, 0), sigmaX=k**j, sigmaY=k**j)
            octave.append(cur_img)
    gaussian_pyramid.append(octave)

2. 计算高斯差分金字塔（DoG）

# 定义高斯差分金字塔层数和尺度
num_octaves = 4
num_scales = 5

# 计算高斯差分金字塔
dog_pyramid = []
for octave in gaussian_pyramid:
    dog = []
    for i in range(num_scales-1):
        cur_img = octave[i+1] - octave[i]
        dog.append(cur_img)
    dog_pyramid.append(dog)

3. 计算拉普拉斯金字塔（LoG）

# 定义拉普拉斯金字塔层数和尺度
num_octaves = 4
num_scales = 5

# 计算拉普拉斯金字塔
laplacian_pyramid = []
for octave in gaussian_pyramid:
    laplacian = []
    for i in range(num_scales-1):
        cur_img = cv2.subtract(octave[i], cv2.pyrUp(octave[i+1]))
        laplacian.append(cur_img)
    laplacian.append(octave[-1])
    laplacian_pyramid.append(laplacian)