您可以使用PIL/pillow创建大的空图像,然后将小图像放在该图像的不同位置。
from PIL import Image
# get images
img1 = Image.open('image1.png')
img2 = Image.open('image2.png')
img3 = Image.open('image3.png')
img4 = Image.open('image4.png')
# get width and height
w1, h1 = img1.size
w2, h2 = img2.size
w3, h3 = img3.size
w4, h4 = img4.size
# to calculate size of new image
w = max(w1, w2, w3, w4)
h = max(h1, h2, h3, h4)
# create big empty image with place for images
new_image = Image.new('RGB', (w*2, h*2))
# put images on new_image
new_image.paste(img1, (0, 0))
new_image.paste(img2, (w, 0))
new_image.paste(img3, (0, h))
new_image.paste(img4, (w, h))
# save it
new_image.save('new.png')
顺便说一句:你可以把它写成for-loop(s)。
与没有 Python 的程序 ImageMagick 相同。
但是您可以在 Python 中通过os.system('convert ...') 使用这些命令
$ convert image1.png image2.png +append row1.png
$ convert image3.png image4.png +append row2.png
$ convert row1.png row2.png -append new.png
+append 加入行,-append 加入列。
即使在一个命令中也可以做到这一点:Stitching Image Set Together
$ convert image1.png image2.png image3.png image4.png +append -crop 2x1@ -append new.png
如果你使用new.pdf 而不是new.png 那么它可以创建PDF
Python 模块Wand 使用ImageMagick。代码类似于pillow。
from wand.image import Image
img1 = Image(filename='image1.png')
img2 = Image(filename='image2.png')
img3 = Image(filename='image3.png')
img4 = Image(filename='image4.png')
w1, h1 = img1.size
w2, h2 = img2.size
w3, h3 = img3.size
w4, h4 = img4.size
w = max(w1, w2, w3, w4)
h = max(h1, h2, h3, h4)
new_image = Image(width=w*2, height=h*2)
new_image.composite(image=img1, left=0, top=0)
new_image.composite(image=img2, left=w, top=0)
new_image.composite(image=img3, left=0, top=h)
new_image.composite(image=img4, left=w, top=h)
new_image.save(filename='new.png')
以cv2 和numpy 为例,但它适用于图像大小相同的情况。如果它们不同,则它们需要空行和空列具有相同的大小。
import cv2
import numpy as np
img1 = cv2.imread('image1.png')
img2 = cv2.imread('image2.png')
img3 = cv2.imread('image3.png')
img4 = cv2.imread('image4.png')
row1 = np.concatenate((img1, img2), axis=1)
row2 = np.concatenate((img3, img4), axis=1)
new_image = np.concatenate((row1, row2))
# or
row1 = np.hstack((img1, img2))
row2 = np.hstack((img3, img4))
new_image = np.vstack((row1, row2))
cv2.imwrite('new.png', new_image)
与matplotlib 和numpy 类似
import matplotlib.image
import numpy as np
img1 = matplotlib.image.imread('image1.png')
img2 = matplotlib.image.imread('image2.png')
img3 = matplotlib.image.imread('image3.png')
img4 = matplotlib.image.imread('image4.png')
row1 = np.concatenate((img1, img2), axis=1)
row2 = np.concatenate((img3, img4), axis=1)
new_image = np.concatenate((row1, row2))
# or
row1 = np.hstack((img1, img2))
row2 = np.hstack((img3, img4))
new_image = np.vstack((row1, row2))
matplotlib.image.imsave('new.png', new_image)
与imageio 和numpy 类似
import imageio
import numpy as np
img1 = imageio.imread('image1.png')
img2 = imageio.imread('image2.png')
img3 = imageio.imread('image3.png')
img4 = imageio.imread('image4.png')
row1 = np.concatenate((img1, img2), axis=1)
row2 = np.concatenate((img3, img4), axis=1)
new_image = np.concatenate((row1, row2))
# or
row1 = np.hstack((img1, img2))
row2 = np.hstack((img3, img4))
new_image = np.vstack((row1, row2))
imageio.imwrite('new.png', new_image)