基于Python+OpenCV的视神经胶质瘤识别：圆内白色像素计算求助

Got it, let's adjust your code to accurately count only the white pixels inside that blue circle you've drawn. The key issues in your current code are:

• You're checking pixels on the image variable (which has the blue circle drawn on it, altering some pixel values) instead of the original image.
• You're iterating over the entire image instead of limiting to the circle's area.

Here are two efficient ways to solve this:

Method 1: Use a Mask (Most Efficient for Large Images) #

This leverages OpenCV's optimized functions to avoid slow Python loops:

import cv2
import numpy as np
import imutils

# Assume your original code up to drawing the circle is here:
# gray = cv2.cvtColor(orig, cv2.COLOR_BGR2GRAY)  # I'm assuming you had this step to get gray
gray = cv2.GaussianBlur(gray, (371, 371), 0)
(minVal, maxVal, minLoc, maxLoc) = cv2.minMaxLoc(gray)
image = orig.copy()
cv2.circle(image, maxLoc, 371, (255, 0, 0), 2)

# Define circle parameters
center = maxLoc
radius = 371
sought = [254, 254, 254]

# Step 1: Create a mask that only keeps the area inside the circle
mask = np.zeros(orig.shape[:2], dtype=np.uint8)
cv2.circle(mask, center, radius, 255, -1)  # -1 fills the circle with white

# Step 2: Isolate white pixels in the original image
white_pixels = cv2.inRange(orig, np.array(sought), np.array(sought))

# Step 3: Keep only white pixels that are inside the circle
circle_white_pixels = cv2.bitwise_and(white_pixels, white_pixels, mask=mask)

# Step 4: Count the non-zero (white) pixels
amount = cv2.countNonZero(circle_white_pixels)
print(f"White pixels inside circle: {amount}")

# Rest of your display code
image = imutils.resize(image, width=400)
cv2.imshow("Optic Image", image)
cv2.waitKey(0)
cv2.destroyAllWindows()

Method 2: Iterate with Circle Boundary Check (For Smaller Images) #

If you prefer to keep the loop approach, fix it to check if pixels are inside the circle and use the original image:

import cv2
import imutils

# Your original setup code
# gray = cv2.cvtColor(orig, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray, (371, 371), 0)
(minVal, maxVal, minLoc, maxLoc) = cv2.minMaxLoc(gray)
image = orig.copy()
cv2.circle(image, maxLoc, 371, (255, 0, 0), 2)

center_x, center_y = maxLoc
radius = 371
radius_sq = radius ** 2  # Precompute square to speed up calculations
sought = (254, 254, 254)
amount = 0

# Iterate over each pixel, but only check those inside the circle
for y in range(orig.shape[0]):
    for x in range(orig.shape[1]):
        # Calculate distance squared to center (avoids sqrt for speed)
        dist_sq = (x - center_x) ** 2 + (y - center_y) ** 2
        if dist_sq <= radius_sq:
            b, g, r = orig[y, x]  # Use original image, not the one with the blue circle!
            if (b, g, r) == sought:
                amount += 1

print(f"White pixels inside circle: {amount}")

# Display code
image = imutils.resize(image, width=400)
cv2.imshow("Optic Image", image)
cv2.waitKey(0)
cv2.destroyAllWindows()

Key Notes: #

• Always use the original image (orig) for pixel checks: The image variable has the blue circle drawn on it, so its pixel values are modified along the circle's edge.
• Mask method is faster: For large medical images, OpenCV's C-optimized functions will outperform Python loops by a huge margin.
• Distance squared trick: Comparing (x - cx)^2 + (y - cy)^2 <= r^2 avoids computing square roots, which saves computation time.

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