使用Python将ND2图像转彩色PNG遇问题，求排查方案

Let's break down what's going wrong here and fix it step by step:

The Root of the Problem #

Your ND2 file's channels are 16-bit single-channel grayscale images (mode 'I'), where the grayscale values represent the brightness of the corresponding color (green channel = green brightness, red channel = red brightness).

Your original code converts this to 8-bit grayscale (convert('L')) and enhances contrast, hence the black-and-white output. When you tried removing convert('L') or forcing an RGB mode, you hit errors because:

• PIL can't automatically map a single-channel grayscale image to an RGB mode (hence the ValueError: image has wrong mode).
• 16-bit values (0-65535) don't scale to 8-bit display ranges by default, leading to all-black images when saved directly.

The Fix: Map Grayscale to RGB Channels #

To retain color, we need to manually map the grayscale data to the corresponding RGB channel (green → G channel, red → R channel) while keeping your existing contrast enhancement and 16-to-8-bit scaling logic.

Modified Code #

Here's the adjusted version of your script with color preservation:

from nd2reader import ND2Reader
from PIL import Image, ImageEnhance
import sys, getopt, os

argv = sys.argv[1:]

def saveImage(imageData, path, color_channel):
    # Convert 16-bit single-channel data to 8-bit grayscale
    im = Image.fromarray(imageData)
    im.mode = 'I'
    # Scale 16-bit values (0-65535) to 8-bit (0-255)
    gray_8bit = im.point(lambda i: i * (1. / 256)).convert('L')
    
    # Create RGB channels based on target color
    if color_channel == 'green':
        r_channel = Image.new('L', gray_8bit.size, 0)
        g_channel = gray_8bit
        b_channel = Image.new('L', gray_8bit.size, 0)
    elif color_channel == 'red':
        r_channel = gray_8bit
        g_channel = Image.new('L', gray_8bit.size, 0)
        b_channel = Image.new('L', gray_8bit.size, 0)
    else:
        # Fallback to grayscale if invalid channel is passed
        rgb_image = gray_8bit
        rgb_image.save(path, "PNG")
        return
    
    # Merge channels into a single RGB image
    rgb_image = Image.merge('RGB', (r_channel, g_channel, b_channel))
    
    # Apply contrast enhancement (keep your original logic)
    contrastEnhancer = ImageEnhance.Contrast(rgb_image)
    im_output = contrastEnhancer.enhance(10)
    im_output.save(path, "PNG")

def convertImage(inputFile):
    fileName = os.path.splitext(inputFile)[0]
    with ND2Reader(inputFile) as images:
        greenFrame = images.get_frame(0)
        redFrame = images.get_frame(1)
        # Pass color channel info to saveImage
        saveImage(greenFrame, fileName + "_green.png", 'green')
        saveImage(redFrame, fileName + "_red.png", 'red')

def main(argv):
    inputfile = ''
    try:
        opts, args = getopt.getopt(argv, "hi:", ["ifile="])
    except getopt.GetoptError:
        print('test.py -i <inputfile>')
        sys.exit(2)
    if(len(opts) < 1):
        print('test.py -i <inputfile>')
        sys.exit(2)
    for opt, arg in opts:
        if opt == '-h':
            print ('test.py -i <inputfile>')
            sys.exit()
        elif opt in ("-i", "--ifile"):
            inputfile = arg
            convertImage(inputfile)

main(argv)

Key Changes Explained #

1. 16-to-8-bit Scaling: We keep your original logic to scale 16-bit values down to 0-255, ensuring the image is visible in standard viewers.
2. RGB Channel Construction: For each color, we create three 8-bit channels: one with the grayscale data (matching the target color) and two filled with 0 (for the other colors).
3. Merge Channels: Image.merge('RGB', ...) combines the three single-channel images into a valid RGB mode image, which PIL can handle without errors.
4. Color-Specific Saving: We pass the target color to saveImage to ensure the right channel is populated.

Optional Optimization #

If your ND2 data uses 12-bit depth instead of 16-bit, adjust the scaling factor to 1. / 4096 instead of 1. / 256 to properly map 0-4095 values to 0-255.

内容的提问来源于stack exchange，提问作者metabolic