Using this script on an image will crop all sub-regions that are surrounded by white pixels boarder as new images. Useful to extract all sub images, say, in a screenshot.

crop_all_sub_pictures.py

#!/usr/bin/env python3

from pathlib import Path

import argparse

from PIL import Image
import numpy as np

from scipy.ndimage.measurements import label


parser = argparse.ArgumentParser(
    "Crop all sub-regions that are surrounded by white pixels"
)

parser.add_argument("image_file")
parser.add_argument("--order-by-y-axis", default=False, action="store_true")
parser.add_argument("--no-region-as-suffix", default=False, action="store_true")
parser.add_argument("--target-pixel-value", default=255)

args = parser.parse_args()


def find_squares(array: np.ndarray, area_range=[-np.inf, np.inf]):
    assert len(array.shape) == 2
    assert array.dtype == bool
    labeled, num_features = label(array, np.ones((3, 3), dtype=int))

    for i in range(1, num_features + 1):
        pixels = np.where(labeled == i)
        # print(pixels)
        x1 = pixels[1].min()
        x2 = pixels[1].max()
        y1 = pixels[0].min()
        y2 = pixels[0].max()
        area = (x2 - x1) * (y2 - y1)
        # print(area)
        print(f"{i} | slice: array[{y1}:{y2}, {x1}:{x2}] (area={area})")

        if area_range[0] <= area <= area_range[1]:
            yield (x1, y1, x2, y2)


def main(args):
    image_fname = Path(args.image_file)
    if not image_fname.exists():
        print(f"Given filename '{image_fname}' does not exists!")
        exit(1)

    image = Image.open(image_fname)
    image_pixels = np.asarray(image)[:, :, :3]

    is_white = (image_pixels == args.target_pixel_value).all(axis=2)

    regions = list(find_squares(~is_white))

    # sort by top-left corners (i.e., [x1, y1])
    # order_by = "order-by-x-axis"
    order_functor = lambda x: (x[0], x[1])

    if args.order_by_y_axis:
        order_functor = lambda x: (x[1], x[0])

    regions = sorted(regions, key=order_functor)

    for i, region in enumerate(regions):
        suffix = ""
        if not args.no_region_as_suffix:
            suffix = f"-[x={region[0]}-{region[2]},y={region[1]}-{region[3]}]"
        target_filename = f"{image_fname.stem}_{i}{suffix}.png"
        image.crop(region).save(target_filename)


if __name__ == "__main__":
    main(args)