如何用Numpy创建环形掩膜？现有代码遇问题求助

Hey there! Let's get your annular mask function working properly. The core issue with your current code is the chained boolean comparison small_radius <= distance_from_center <= big_radius—NumPy doesn't handle Python's native chained comparison logic for arrays, which will lead to wrong mask results or broadcasting errors.

Step 1: Fix the Boolean Condition #

Instead of the chained comparison, you need to explicitly combine two separate conditions using np.logical_and (or the & operator, with parentheses to avoid precedence issues).

Step 2: Handle Pixel-to-Unit Conversion #

Since you mentioned each pixel corresponds to 4 units, we should add logic to convert your input radii and center coordinates from "units" to pixel values (unless you're already passing pixel-based values directly). This ensures the mask aligns correctly with your 256×256 pixel data.

Corrected Function #

Here's the revised version of your function with these fixes, plus a flexible parameter for pixel-unit conversion:

import numpy as np

def createAnnularMask(dimx, dimy, center, big_radius, small_radius, pixels_per_unit=4):
    # Convert input units to pixel coordinates/radii
    center_pixel = (center[0] / pixels_per_unit, center[1] / pixels_per_unit)
    big_radius_pixel = big_radius / pixels_per_unit
    small_radius_pixel = small_radius / pixels_per_unit
    
    # Create grid of coordinates
    Y, X = np.ogrid[:dimx, :dimy]
    
    # Calculate distance from each pixel to the center
    distance_from_center = np.sqrt((X - center_pixel[0])**2 + (Y - center_pixel[1])**2)
    
    # Create the annular mask using valid boolean combination
    mask = np.logical_and(
        distance_from_center >= small_radius_pixel,
        distance_from_center <= big_radius_pixel
    )
    
    # Alternative using & operator (note the parentheses):
    # mask = (distance_from_center >= small_radius_pixel) & (distance_from_center <= big_radius_pixel)
    
    return mask

How to Use It #

If your input parameters are in units (e.g., center at (512, 512) units, outer radius 100 units, inner radius 50 units for a 256×256 pixel grid):

# Generate mask for 256x256 data
mask = createAnnularMask(
    dimx=256, dimy=256,
    center=(512, 512),
    big_radius=100,
    small_radius=50
)

# Optional: Visualize the mask
import matplotlib.pyplot as plt
plt.imshow(mask, cmap='gray')
plt.title("Annular Mask")
plt.show()

If you're already passing pixel-based values, just set pixels_per_unit=1:

mask = createAnnularMask(
    dimx=256, dimy=256,
    center=(128, 128),  # Pixel center
    big_radius=25,      # Pixel radius
    small_radius=12,
    pixels_per_unit=1
)

Key Notes #

• The np.ogrid creates a grid of coordinates that efficiently calculates distances without generating a full 2D array for X and Y (saves memory).
• Using np.logical_and ensures the boolean operations work correctly across the entire NumPy array.
• The pixels_per_unit parameter makes the function flexible if your pixel-unit ratio changes later.

内容的提问来源于stack exchange，提问作者bhjghjh