如何用Python-OpenCV提取图像绿色圆形区域？原代码失效求助

Hey there! The issue with your current code is that it's using grayscale thresholding and contour detection without leveraging the green color of your target circle—so it might be picking up other high-contrast regions instead of just the green area. Let's adjust the approach to focus specifically on isolating the green circle, which will give you the clean image B you need.

Step-by-Step Solution #

We'll use color segmentation in the HSV color space (it's far more reliable for color-based isolation than BGR) combined with simple morphological operations to clean up noise.

1. Full Working Code #

import cv2
import numpy as np
import os

# Load your image A (the one with the green circle)
circle_path_test = r'D:\rec.png'
img = cv2.imread(circle_path_test)
if img is None:
    print("Error: Could not load image!")
    exit()

# Convert to HSV color space (better for color segmentation)
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)

# Define HSV range for green (adjust these values if your green is lighter/darker)
# These are typical ranges—tweak them using an HSV color picker if needed
lower_green = np.array([40, 40, 40])
upper_green = np.array([70, 255, 255])

# Create a mask for green regions
mask = cv2.inRange(hsv, lower_green, upper_green)

# Clean up the mask: remove small noise and fill gaps
# Use morphological operations (dilation followed by erosion)
kernel = np.ones((5, 5), np.uint8)
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel)
mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)

# Optional: If you already have circle coordinates from cv2.HoughCircles, use this to refine the mask
# Example (replace with your HoughCircles results):
# circles = cv2.HoughCircles(gray, cv2.HOUGH_GRADIENT, 1, 20, param1=50, param2=30, minRadius=0, maxRadius=0)
# if circles is not None:
#     circles = np.uint16(np.around(circles))
#     mask = np.zeros(img.shape[:2], np.uint8)
#     for i in circles[0, :]:
#         cv2.circle(mask, (i[0], i[1]), i[2], 255, -1)

# Apply the mask to keep only the green circle (set everything else to black)
img_b = cv2.bitwise_and(img, img, mask=mask)

# Save the result
save_path = os.path.join(os.getcwd(), 'isolated_green_circle.png')
cv2.imwrite(save_path, img_b)

# Optional: Preview the result
# cv2.imshow("Isolated Green Circle", img_b)
# cv2.waitKey(0)
# cv2.destroyAllWindows()

2. Key Improvements Over Your Original Code #

• Color-Based Segmentation: Instead of grayscale thresholding, we target the green color directly in HSV, which ensures we only focus on the circle's unique color signature.
• Mask Cleaning: Morphological operations (MORPH_CLOSE fills small gaps in the green region, MORPH_OPEN removes tiny noise spots) make the mask cleaner and more accurate.
• Optional HoughCircles Integration: If you already have the circle's center and radius from cv2.HoughCircles, you can draw a perfect circle mask instead of relying on color alone—this is even more precise if the green circle has uniform color.

3. How to Adjust the Green HSV Range #

If the mask isn't capturing your green circle perfectly:

1. Use an HSV color picker tool to get the exact HSV values of your green circle.
2. Adjust lower_green and upper_green accordingly. For example, if your green is very bright, lower the saturation threshold; if it's dark, lower the value threshold.

内容的提问来源于stack exchange，提问作者user9077223