Save numpy array as single channel png with custom colors

问题

I have an integer numpy array representing image with few values (about 2-5). And I would like to save it to png file with custom color for every value. I was trying it like this:

import numpy as np
from PIL import Image

array = np.zeros([100, 200, 4], dtype=np.uint8)
array[:,:100] = [255, 128, 0, 255] #Orange left side
array[:,100:] = [0, 0, 255, 255]   #Blue right side

img = Image.fromarray(array)
img.save(r'D:\test.png')

The result is ok, but it has 4 channels. I need the result to be single channel with custom colors.

I was trying it this way:

array = np.zeros([100, 200], dtype=np.uint8)
array[:,:100] = 1 
array[:,100:] = 0   

the result is single channel, but it is of course graysale. I can't figure out how to assign color to values 1 and 0 and save it as single channel. Was trying play around with matplotlib colormaps, but with no success.

Any help will be very appreciated

回答1:

You can make a palette image like this:

#!/usr/bin/env python3

from PIL import Image
import numpy as np

# Make image with small random numbers
im = np.random.randint(0,5, (4,8), dtype=np.uint8)

# Make a palette
palette = [255,0,0,    # 0=red
           0,255,0,    # 1=green
           0,0,255,    # 2=blue
           255,255,0,  # 3=yellow
           0,255,255]  # 4=cyan
# Pad with zeroes to 768 values, i.e. 256 RGB colours
palette = palette + [0]*(768-len(palette))

# Convert Numpy array to palette image
pi = Image.fromarray(im,'P')

# Put the palette in
pi.putpalette(palette)

# Display and save
pi.show()
pi.save('result.png')

来源：https://stackoverflow.com/questions/64681931/save-numpy-array-as-single-channel-png-with-custom-colors