Batch overlay/merge images with similar filenames

2
votes

I have 1000+ images in a directory labelled:
x_1.bmp
x_2.bmp
x_3.bmp
y_1.bmp
y_3.bmp
...etc

Ideally, I'd like to overlay all files with a similar name and export as x_P.png, y_P.png, etc. as below. All images are the same size and none of the colour pixels overlap.

x_1 example + x_2 example + x_3 example = enter image description here

I've tried to batch overlay images using a merge-divide script function in GIMP/BIMP without success. Would PIL.Image.Image.paste() be able to do this, or something similar? I'm using Windows and Python 3.7.3

Thanks

2
pythonimagepython-imaging-librarygimp
So, you want to convert all files from .bmp to .png? What do you mean by "overly all files with a similar name"? – Ayush Garg
Yes, I'll need to save the final image as a png but this can be done separately if necessary. All files starting with x should be merged into one image, all files starting with y should be merged into a separate image. All images are the same size. – user13024726
What do you mean by "merge"? Are the images going to be layered on top of each other, placed next to each other, etc. – Ayush Garg
Layered on top of each other – user13024726
Is x_1.bmp the red channel and x_2.bmp the green channel? Please show your images and the expected result, else we are all guessing and wasting our time. Thank you. – Mark Setchell

2 Answers

0
votes

Working off the example from the other answer I managed to sort this and maintain the RGB values.

import os
from PIL import Image
import PIL
import itertools as it


path = path/to/files
alpha = 1
list_of_imgs = sorted(os.listdir('path'))
for k,g in it.groupby(list_of_imgs, key=lambda x: x.split('_')[0]): 
    img = Image.open('path/{}'.format(next(g)),'r').convert('RGBA') 
    x,y = img.size 
    g_list = list(g) 
    for i,item in enumerate(g_list): 
        next_img = Image.open('path/{}'.format(item)).convert('RGBA')
        if i == 0: 
            new_img = Image.alpha_composite(next_img,img) 
        else: 
            new_img = Image.alpha_composite(next_img,new_img) 
        next_img.paste(new_img, (0,0,x,y), mask=new_img) 
        new_img.save(f'path/{k}_P.png',format='png') 


Image.open(path + r"/output_P.png").getcolors()

Result

Result

0
votes

I was able to do this for a small number of images. I'm not sure what it would accomplish with a lot of images, though, since everything would jumble together. Anyway, let me know if it helps. I am using pillow version 7.0.0 here.

import os
import itertools as it
from PIL import Image

alpha = 0.5
list_of_imgs = sorted(os.listdir('images'))
for k,g in it.groupby(l,key=lambda x: x.split('_')[0]): 
    img = Image.open('images/{}'.format(next(g)),'r').convert('RGBA') 
    x,y = img.size 
    g_list = list(g) 
    for i,item in enumerate(g_list): 
        next_img = Image.open('images/{}'.format(item)).convert('RGBA')
        if i == 0: 
            new_img = Image.blend(next_img,img,alpha) 
        else: 
            new_img = Image.blend(next_img,new_img,alpha) 
        next_img.paste(new_img, (0,0,x,y), mask=new_img) 
        new_img.save(f'images/{k}.png',format='png')

Result for three images (dog/puppy, mountain peak, lighthouse): enter image description here