如何通过中心坐标、宽高及角度绘制带填充值的旋转矩形？

First off, I feel your pain—having to cobble together rotation logic every time you need a rotated filled rectangle is annoying, and none of the major libraries (OpenCV, PIL, Matplotlib) have that exact one-liner you're asking for. But the good news is we can easily wrap the necessary steps into a custom function that matches your desired clean API.

Let's Build the Wrapper Function #

Below are practical implementations for both OpenCV and PIL/Pillow, the most common libraries for image drawing tasks.

OpenCV Implementation #

OpenCV handles geometric transformations smoothly, and we can skip full image warping by calculating the rotated rectangle's vertices directly to fill it:

import cv2
import numpy as np

def draw_rotated_filled_rectangle(img, centerx, centery, width, height, angle, value):
    # Convert angle to radians for math calculations
    angle_rad = np.deg2rad(angle)
    half_w = width / 2
    half_h = height / 2

    # Define rectangle corners relative to the center (0,0)
    corners = np.array([
        [-half_w, -half_h], [half_w, -half_h],
        [half_w, half_h], [-half_w, half_h]
    ], dtype=np.float32)

    # Create rotation matrix
    rotation_matrix = np.array([
        [np.cos(angle_rad), -np.sin(angle_rad)],
        [np.sin(angle_rad), np.cos(angle_rad)]
    ])

    # Rotate corners and translate to the target center coordinates
    rotated_corners = np.dot(corners, rotation_matrix) + np.array([centerx, centery])
    rotated_corners = rotated_corners.astype(np.int32).reshape((-1, 1, 2))

    # Fill the rotated rectangle polygon
    cv2.fillPoly(img, [rotated_corners], color=value)
    return img

How to Use It

# Create a blank 500x500 grayscale image
img = np.zeros((500, 500), dtype=np.uint8)

# Draw your desired rotated rectangle: center (250,250), width 100, height 50, angle 30°, fill value 5
img = draw_rotated_filled_rectangle(img, 250, 250, 100, 50, 30, 5)

# View the result
cv2.imshow("Rotated Filled Rectangle", img)
cv2.waitKey(0)
cv2.destroyAllWindows()

PIL/Pillow Implementation #

For PIL users, we can calculate rotated vertices and use ImageDraw.polygon() to fill the shape:

from PIL import Image, ImageDraw
import math

def draw_rotated_filled_rectangle(img, centerx, centery, width, height, angle, value):
    draw = ImageDraw.Draw(img)
    half_w = width / 2
    half_h = height / 2
    angle_rad = math.radians(angle)

    # Define corners relative to the center
    corners = [
        (-half_w, -half_h), (half_w, -half_h),
        (half_w, half_h), (-half_w, half_h)
    ]

    # Rotate each corner and translate to target center
    rotated_corners = []
    for x, y in corners:
        rotated_x = x * math.cos(angle_rad) - y * math.sin(angle_rad) + centerx
        rotated_y = x * math.sin(angle_rad) + y * math.cos(angle_rad) + centery
        rotated_corners.append((rotated_x, rotated_y))

    # Fill the polygon with your desired value
    draw.polygon(rotated_corners, fill=value)
    return img

How to Use the PIL Version

# Create a blank 500x500 grayscale image
img = Image.new("L", (500, 500), 0)

# Draw the rotated rectangle
img = draw_rotated_filled_rectangle(img, 250, 250, 100, 50, 30, 5)

# View or save the result
img.show()

Key Notes #

• Both functions mimic the simple API you wanted, with the standard addition of passing the target image as the first parameter (this keeps it consistent with how drawing works in most libraries).
• The value parameter works for grayscale images (like your example with 5); for color images, pass a tuple such as (255, 0, 0) for red.
• If you strictly want the draw.rectangle(...) syntax you mentioned, you could wrap this into a class that holds a reference to your image—though the above functions are more flexible for most use cases.

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