如何使用 python 中的颜色通道分离图像？

Question

Hello I want to separate an object based it's color into a new image using python PIL

but I don't quite get the gasp of it. I'm still new with python.

what I already done is a code from matlab and it works

G = imread('x.png');
L = G;
for i=1:217 #this is image height
    for j=1:286 #this is image width
        for k=1:3 #this is color RGB channel
            if(L(i,j,2) < 174) #if green value is under 174
                L(i,j,k) = L(i,j,k);
            else
                L(i,j,k) = 256; #change to white
            end

            if(L(i,j,3) < 174) #if blue value is under 174
                L(i,j,k) = L(i,j,k);
            else
                L(i,j,k) = 256; #change to white
            end
        end
    end
end
imshow(L)

Can I get a proper explanation about how you do it using python PIL?

thank you very much

edit:

What I'm trying to do is something like this

https://imgur.com/6n1QCfR

Then the result is this

https://imgur.com/p99kN5p

Answer 1

Here is a very easy way to do it using PIL/Pillow:

#!/usr/bin/env python3

from PIL import Image
import numpy as np

# Load image and ensure it is RGB (not palette)
im = Image.open('circles.jpg').convert('RGB')

# Make into Numpy array so we can do fast, vectorised operations
na = np.array(im)

# Make a mask, which is True wherever Blue channel is high, False elsewhere - see #1 below
mBlueHi = na[...,2] >= 174

# Likewise for Green channel - see #2 below
mGreenHi = na[...,1] >= 174

# In our original image "na", anywhere the Green or Blue mask is set, make it white
na[mGreenHi | mBlueHi] = [255,255,255]

# Make back into PIL Image and save
Image.fromarray(na).save('result.png')

Note #1: The image is stored in a 3 dimensional array, height x width x RGB channels. Red is in channel 0, so it is na[:,:,0] and na[...,0] is shorthand for that same thing.

Note #2: Green is in the second channel starting from index 0, so the green pixels can be addressed using na[:,:,1] or equally na[...,1]

Note that JPEG is lossy and often a poor choice for intermediate files in image processing because it changes values to make files smaller. Consider using PNG which is lossless.