使用NumPy与Matplotlib绘制RGB矩阵时颜色显示异常的问题排查

Hey there! Let's walk through the issues with your code and get that flag looking right.

Key Problems in Your Current Code #

• Mismatched color value range and data type:
  Matplotlib's imshow() treats floating-point arrays as having values in the range 0.0 to 1.0. Your code creates a float array (the default for np.empty()) but assigns RGB values from 0 to 255—all these values get clamped to 1.0, which is why your colors are messed up. For 0-255 RGB values, you need an integer array of type uint8 (8-bit unsigned integer).

• Tiny image dimensions:
  Your array is (1, 3, 3)—that's a single pixel tall and 3 pixels wide. Even if the colors were correct, you'd barely be able to see anything! Belgian flags are vertical stripes, so you want a taller, wider canvas to display them properly.

Fixed Code Example #

Here's adjusted code that addresses both issues and renders the Belgian flag correctly:

import numpy as np
import matplotlib.pyplot as plt

# Create a properly sized RGB array with uint8 type (for 0-255 RGB values)
flag_height = 200
flag_width = 600
flag = np.zeros((flag_height, flag_width, 3), dtype=np.uint8)

# Split the width into 3 equal vertical blocks for each color
block_width = flag_width // 3

# Assign colors to each stripe (black, yellow, red)
flag[:, :block_width, :] = (0, 0, 0)
flag[:, block_width:2*block_width, :] = (254, 240, 71)
flag[:, 2*block_width:, :] = (255, 55, 14)

# Display the flag without axes for a cleaner look
plt.imshow(flag)
plt.axis('off')
plt.show()

What Changed? #

1. Data type set to uint8:
   Using dtype=np.uint8 tells NumPy and Matplotlib that we're working with standard 0-255 RGB values, so no clamping occurs.
2. Resized the flag:
   A 200x600 canvas makes the vertical stripes clearly visible.
3. Proper stripe layout:
   We split the width into three equal sections to match the Belgian flag's vertical stripe pattern.

Bonus: Using Floating-Point Arrays (If You Prefer) #

If you want to stick with floating-point values, just divide your RGB values by 255 to fit the 0.0-1.0 range:

flag = np.zeros((200, 600, 3), dtype=np.float32)
flag[:, :block_width, :] = (0/255, 0/255, 0/255)
flag[:, block_width:2*block_width, :] = (254/255, 240/255, 71/255)
flag[:, 2*block_width:, :] = (255/255, 55/255, 14/255)

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