要求：将图片裁剪成九宫格，宫格图片位置打乱

思路：

判断该图片长宽是否能被3整除，不能需要对图片进行裁剪（我的是边角裁剪）
双重for循环，根据步长对图片数组进行切片达到裁剪效果
random.shuffle()达到图片位置打乱的效果
使用np.concatenate()对图片数组按维拼接

代码：

第三方库

import numpy as np
from PIL import Image
import matplotlib.pyplot as plt
import random
import math

判断函数

def judge(img, n=9):
    x = int(math.sqrt(n))
    w, h, _ = img.shape

    if w % x:
        w = (w // x) * x
    if h % x:
        h = (h // x) * x
    
    return w, h

裁剪拼接函数

def crop_nine(new_image):
    shape_w, shape_h, _ = new_image.shape
    patch_w = shape_w // 3
    patch_h = shape_h // 3
    patch = []
    for w in range(0, shape_w, patch_w):
        for h in range(0, shape_h, patch_h):
            patch.append(new_image[w:w+patch_w, h:h+patch_h, :])
    random.shuffle(patch)

    img_1 = np.concatenate((patch[0], patch[1], patch[2]), axis=0)
    img_2 = np.concatenate((patch[3], patch[4], patch[5]), axis=0)
    img_3 = np.concatenate((patch[6], patch[7], patch[8]), axis=0)
    img = np.concatenate((img_1, img_2, img_3), axis=1)
    return img

运行部分：

image_path = "dog.jpg"
img = np.array(Image.open(image_path))
w, h = judge(img)
new_image = img[:w, :h, :]
new_image = crop_nine(new_image)
plt.imshow(new_image)