如何在PyQt的QLabel中将numpy灰度图像转为伪彩色显示？

Hey there! You're already halfway there with displaying grayscale images in QLabel—adding pseudocolor just requires a couple extra steps to map your grayscale values to a color palette, similar to how matplotlib's imshow works with colormaps. Here's how to tweak your code to make it happen:

Key Steps to Add Pseudocolor #

• First, normalize your grayscale array to the range [0, 1] (most colormaps expect this input range)
• Apply a colormap to convert grayscale values to RGB/A color data
• Convert the resulting color array to a QImage for PyQt to render

Modified Code with Pseudocolor Support #

We'll use matplotlib's built-in colormaps (like viridis, jet, etc.) since they're easy to work with, but you could also define a custom colormap if needed. Here's your updated code:

import sys
from PyQt5.QtCore import *
from PyQt5.QtGui import *
from PyQt5.QtWidgets import *
import numpy as np
import qimage2ndarray
import matplotlib.cm as cm  # Import matplotlib's colormap module

class MyLabel(QLabel):
    def __init__(self):
        super(MyLabel, self).__init__()
        img = np.zeros((256,256))
        img[0:128,0:128] = 255
        self.img = img

    def paintEvent(self, QPaintEvent):
        super(MyLabel, self).paintEvent(QPaintEvent)
        
        # Step 1: Normalize grayscale image to [0, 1] range
        normalized_img = self.img / self.img.max()  # Handles any value range, not just 0-255
        
        # Step 2: Apply colormap (replace 'viridis' with any matplotlib colormap name)
        colormap = cm.get_cmap('viridis')
        pseudocolor_img = colormap(normalized_img)  # Outputs RGBA array (shape: (256,256,4))
        
        # Step 3: Convert RGBA array to QImage (array2qimage handles color arrays)
        QImg = qimage2ndarray.array2qimage(pseudocolor_img)
        
        pos = QPoint(0, 0)
        source = QRect(0, 0, 256,256)
        painter = QPainter(self)
        painter.drawPixmap(pos, QPixmap.fromImage(QImg), source)

class Window(QWidget):
    def __init__(self):
        super(Window, self).__init__()
        layout = QHBoxLayout(self)
        self.label = MyLabel()
        layout.addWidget(self.label)

if __name__ == '__main__':
    app = QApplication(sys.argv)
    window = Window()
    window.show()
    sys.exit(app.exec_())

What Changed? #

• Added import matplotlib.cm as cm to access pre-built colormaps
• Normalization: We scale the grayscale values to [0,1] so the colormap applies correctly (this works even if your image has values outside the 0-255 range)
• Colormap Application: cm.get_cmap('viridis') loads your chosen palette, and calling it on the normalized array returns an RGBA color array
• QImage Conversion: Instead of gray2qimage, we use array2qimage since we now have a 4-channel (RGBA) color array

Custom Colormap Option (No Matplotlib Dependency) #

If you don't want to rely on matplotlib, you can create a simple colormap manually. For example, a blue-to-red gradient:

def custom_colormap(gray_value):
    # gray_value is normalized to [0,1]
    r = gray_value
    b = 1 - gray_value
    g = 0.0
    return (r, g, b, 1.0)  # RGBA

# Replace the colormap section in paintEvent with:
pseudocolor_img = np.apply_along_axis(custom_colormap, 2, normalized_img[..., np.newaxis])

This will give you a red-blue gradient without needing matplotlib installed.

内容的提问来源于stack exchange，提问作者Qiang Zhang