Python 判断图片相似度的方法

这里是博主自己使用不同的方法对相似度的测试

最后还是觉得最后一个方法的准确度比较高，其有效的降低了图片尺寸不一致导致的不准确性

import cv2
import imagehash as imagehash
import numpy as np
from PIL import Image
from skimage import io, transform
from skimage.metrics import structural_similarity as ssim
# image1 = cv2.imread(r'D:\BusinessProject\desktop-program\test_test\IMG\1.jpg')
# image2 = cv2.imread(r'D:\BusinessProject\desktop-program\test_test\IMG\2.jpg')

# image1 = cv2.resize(image1, (400, 400))
# image2 = cv2.resize(image2, (400, 400))
#
# gray1 = cv2.cvtColor(image1, cv2.COLOR_BGR2GRAY)
# gray2 = cv2.cvtColor(image2, cv2.COLOR_BGR2GRAY)
#
# ssim_score = cv2.compareSSIM(gray1, gray2)
# print("SSIM Score:", ssim_score)

# image1 = io.imread(r'D:\BusinessProject\desktop-program\test_test\IMG\1.jpg')
# image2 = io.imread(r'D:\BusinessProject\desktop-program\test_test\IMG\2.jpg')
#
# # 计算图像尺寸
# image1_height, image1_width = image1.shape[:2]
# image2_height, image2_width = image2.shape[:2]
#
# # 设置窗口大小为图像中较小的一侧
# win_size = min(image1_height, image1_width, image2_height, image2_width)
#
# # 计算两个图片的相似度
# similarity = ssim(
#     image1,
#     image2,
#     win_size=min(win_size, 7),  # 确保窗口大小小于等于较小的一侧，并至少为 7
#     multichannel=True
# )
#
# # 打印相似度
# print("图片相似度：", similarity)


# image1 = Image.open(r'D:\BusinessProject\desktop-program\test_test\IMG\1.jpg')
# image2 = Image.open(r'D:\BusinessProject\desktop-program\test_test\IMG\4(1).jpg')
# phash1 = imagehash.phash(image1)
# phash2 = imagehash.phash(image2)
# hamming_distance = phash1 - phash2
# print("相似度得分: ", 1 - (hamming_distance / len(phash1.hash)) )



image1 = cv2.imread(r'D:\BusinessProject\desktop-program\test_test\IMG\1.jpg', cv2.IMREAD_GRAYSCALE)
image2 = cv2.imread(r'D:\BusinessProject\desktop-program\test_test\IMG\4(1).jpg', cv2.IMREAD_GRAYSCALE)
orb = cv2.ORB_create()
keypoints1, descriptors1 = orb.detectAndCompute(image1, None)
keypoints2, descriptors2 = orb.detectAndCompute(image2, None)

bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)
matches = bf.match(descriptors1, descriptors2)
matches = sorted(matches, key=lambda x: x.distance)
similarity_score = len(matches) / len(keypoints1) * 100
print(similarity_score)