python 将图片crop成指定大小的

这种方法是均匀将图片分成50*50大小的。没有重叠部分。

from PIL import Image
import os


def image_cut(path_name, id, box_w, box_h):

    # path_name:要剪切的图像的文件夹(路径)
    # id:图像的命名序号（按我提供的重命名方式命名的序号，参考：图片批量重命名）
    # box_w：要剪切的图形的宽度
    # box_h：要剪切的图形的高度

    name = os.path.join(path_name, str(id) + ".png")  # 要剪切图像的名字
    im = Image.open(name)
    # im_size = im.size # 像素大小
    width = im.size[0]
    height = im.size[1]

    if width > box_w & (height > box_h):
        for i in range(int(width / box_w)):
            for j in range(int(height / box_h)):
                name_cut = os.path.join(path_name, 'img_' + str(id) + '_' + str(i) + '_' + str(j) + ".jpg")
                cm = im.crop(box=(i * box_w, j * box_h, (i + 1) * box_w, (j + 1) * box_h))
                cm.save(name_cut)
                if i + 1 == int(width / box_w):
                    name_cut = os.path.join(path_name, 'img_' + str(id) + '_' + str(i + 1) + '_' + str(j) + ".jpg")
                    cm = im.crop(box=((i + 1) * box_w, (j * box_h), width, (j + 1) * box_h))
                    cm.save(name_cut)
                if j + 1 == int(height / box_h):
                    name_cut = os.path.join(path_name, 'img_' + str(id) + '_' + str(i) + '_' + str(j + 1) + ".jpg")
                    cm = im.crop(box=(i * box_w, (j + 1) * box_h, (i + 1) * box_w, height))
                    cm.save(name_cut)
                if i + 1 == int(width / box_w) & j + 1 == int(height / box_h):
                    name_cut = os.path.join(path_name, 'img_' + str(id) + '_' + str(i + 1) + '_' + str(j + 1) + ".jpg")
                    cm = im.crop(box=((i + 1) * box_w, (j + 1) * box_h, width, height))
                    cm.save(name_cut)


path_name = "/home/li/data/data1(crop)/train/aug_gt_scale_0.9/" # 要剪切的图像的文件夹路径

for item in os.listdir(path_name):
    id=item.split('.')[0]
    image_cut(path_name,id,box_w=50,box_h=50)