如何将 24 帧编码为一张 24 位 RGB 图像？

Question

I want to add Blender support for a light field display device. The way the display works with standard HDMI cable is the video has each 24bit RGB frame store 24 monochrome frames. The display splits those frames as it receives them. Here's an illustration:

For this to work, of course, the monochrome frames must be properly encoded in a single 24bit RGB image first. Can PIL, Pillow or another Python library allow to do this?

Answer 1

Not production quality, but probably enough to get you started, I think it must be something like this:

#!/usr/bin/env python3

from PIL import Image
import numpy as np

w, h = 192, 192

# Create RGB output array h, w, 3
out = np.zeros((h,w,3),dtype=np.uint8)

file = 1
# Iterate over 3 channels, Red, Green and Blue- or maybe B, G, R
for channel in range(3):
    # Iterate over the 8 bits in each channel
    for bitpos in range(8): 
        filename = f"{file}.png"
        print(f"Opening {filename} as bit {bitpos} of channel {channel}")
        im = Image.open(filename).convert('L') 

        # Convert to Numpy image with range 0..1
        ni = (np.array(im)/255).astype(np.uint8)

        # Shift and stuff into output array
        out[...,2-channel] |= ni << (bitpos) 
        file += 1

result = Image.fromarray(out)
result.save('result.png')

Sample Output

Opening 1.png as bit 0 of channel 0
Opening 2.png as bit 1 of channel 0
Opening 3.png as bit 2 of channel 0
Opening 4.png as bit 3 of channel 0
Opening 5.png as bit 4 of channel 0
Opening 6.png as bit 5 of channel 0
Opening 7.png as bit 6 of channel 0
Opening 8.png as bit 7 of channel 0
Opening 9.png as bit 0 of channel 1
Opening 10.png as bit 1 of channel 1
Opening 11.png as bit 2 of channel 1
Opening 12.png as bit 3 of channel 1
Opening 13.png as bit 4 of channel 1
Opening 14.png as bit 5 of channel 1
Opening 15.png as bit 6 of channel 1
Opening 16.png as bit 7 of channel 1
Opening 17.png as bit 0 of channel 2
Opening 18.png as bit 1 of channel 2
Opening 19.png as bit 2 of channel 2
Opening 20.png as bit 3 of channel 2
Opening 21.png as bit 4 of channel 2
Opening 22.png as bit 5 of channel 2
Opening 23.png as bit 6 of channel 2
Opening 24.png as bit 7 of channel 2