Python code using PIL to split the image into 4 quarters. useful for the four-view image splitter 3D-PTV project on http://www.openptv.net

split_image_4_quarters.py

from PIL import Image
import os
import glob
import numpy as np


def crop(im, height, width):
    # im = Image.open(infile)
    imgwidth, imgheight = im.size
    rows = np.int(imgheight/height)
    cols = np.int(imgwidth/width)
    for i in range(rows):
        for j in range(cols):
            # print (i,j)
            box = (j*width, i*height, (j+1)*width, (i+1)*height)
            yield im.crop(box)


if __name__ == '__main__':
    # change the path and the base name of the image files
    imgdir = '.'
    basename = 'Cam-*.tif'
    filelist = glob.glob(os.path.join(imgdir, basename))
    for filenum, infile in enumerate(filelist):
        # infile='/Users/alex/Documents/PTV/test_splitter/cal/Camera 1-1-9.tif'
        print(f'file no = {filenum}')  # keep the numbers for the future
        print(f'name is {infile}')
        im = Image.open(infile)
        imgwidth, imgheight = im.size
        print(('Image size is: %d x %d ' % (imgwidth, imgheight)))
        height = np.int(imgheight/2)
        width = np.int(imgwidth/2)
        start_num = 0
        for k, piece in enumerate(crop(im, height, width), start_num):
            # print k
            # print piece
            img = Image.new('L', (width, height), 255)
            # print img
            img.paste(piece)
            path = os.path.join("cam%d_1%05d.tif" % (int(k+1),filenum))
            img.save(path)
            os.rename(path, os.path.join("cam%d.1%05d" % (int(k+1), filenum)))

This gives images in grayscale , what if I need colored images

@linux-devil, just replace img=Image.new('L', (width,height), 255) with img=Image.new('RGB', (width,height), 255). You will get the color images.

do you have code for stitching it back together?

do you have code for stitching it back together?

@atzin-em Not sure why would you like to stick back together - the original image im is the one which is full size already. Just use it.

I need to edit and replace certain parts of the split up image and then put it back together.

I need to edit and replace certain parts of the split up image and then put it back together.

this is usually done on the image itself, no need to split it into separate files, just use slicing and numpy. For instance:

I = numpy.asarray(PIL.Image.open('test.jpg'))
 I[:512,:512] *= 2
im = PIL.Image.fromarray(numpy.uint8(I))

what if I want to split it into [ m ] by [ n ] what do I have to change?

https://scikit-image.org/docs/dev/auto_examples/numpy_operations/plot_view_as_blocks.html#sphx-glr-auto-examples-numpy-operations-plot-view-as-blocks-py

img=Image.new('RGB', (int(width),int(height)), 255)
Shouldn't we convert the width and height to integer type? Without that, it's showing an error.
Image.new function accepts the pixel sizes so I think it has to be integers.

Thanks for the message. I have updated the code. Please:

1. use Python 3 (preferable Conda distribution)
2. Install numpy, pillow (replaces PIL)
3. run with a demo file included Cam-0001.tif

you should get 4 images cam1.10000, cam2.10000 and so on.

Cam1-0001.tif

img=Image.new('RGB', (int(width),int(height)), 255)
Shouldn't we convert the width and height to integer type? Without that, it's showing an error.
Image.new function accepts the pixel sizes so I think it has to be integers.

please send a link to your image - I cannot debug it with my images.

img=Image.new('RGB', (int(width),int(height)), 255)
Shouldn't we convert the width and height to integer type? Without that, it's showing an error.
Image.new function accepts the pixel sizes so I think it has to be integers.

please send a link to your image - I cannot debug it with my images.
your code works fine on my image. No issues, I got 4 images from the original one. I just wanted to notify you about that error. Thanks a lot for the code.

Thanks. in the new version I think I fixed it with

        height = np.int(imgheight/2)
        width = np.int(imgwidth/2)