检测图像中间的颜色（OpenCV，Python）

Question

I have these 3 images (consider "image" the whole square not only the figure inside - this is just for demonstration purposes):

What I want to do is detect the colour in the middle (center) of each one. So, having an area (square or circle) in the center and with OpenCV detect which is the colour. Something like a color picker...

The purpose is to have 3 values, 3 for each image (BGR).

Example:

What is the colour in the ROI ?

Thanks

EDIT

Using this code I can find the middle of an image and apply a mask.

import cv2
import numpy as np

img = cv2.imread("im2.png")

height, width, depth = img.shape
circle_img = np.zeros((height, width), np.uint8)

mask = cv2.circle(circle_img, (int(width / 2), int(height / 2)), 20, 1, thickness=-1)
masked_img = cv2.bitwise_and(img, img, mask=circle_img)

cv2.imshow("masked", masked_img)
cv2.waitKey(0)

Now it remains to find BGR of the masked area (visible one...)

Answer 1

import cv2
import numpy as np

img = cv2.imread('image.png')
height, width = img.shape[:2]
# Change these values to fit the size of your region of interest
roi_size = 10 # (10x10)
roi_values = img[(height-roi_size)/2:(height+roi_size)/2,(width-roi_size)/2:(width+roi_size)/2]
mean_blue = np.mean(roi_values[:,:,0])
mean_green = np.mean(roi_values[:,:,1])
mean_red = np.mean(roi_values[:,:,2])

print("R: {}  G: {}  B: {}").format(mean_red, mean_green, mean_blue)  

Answer 2

确保图像是PhotoImage ，计算中间点并获取颜色：

r,g,b = image.get(x, y)

Answer 3

Using Pillow:

from PIL import Image

im = Image.open(path_to_image)

width, height = im.size #get image size

#since you do not specify the size of the center area in the question, just replace the below variables with what you want
left = (width - your_width)/2
top = (height - your_height)/2
right = (width + your_width)/2
bottom = (height + your_height)/2

im.crop((left, top, right, bottom)) #crop the center of the image

rgb = im.convert('RGB') # get three R G B values
r, g, b = rgb.getpixel((1, 1))

print(r,g,b)

With OpenCV replace the two first lines with:

im = cv2.imread(path_to_image)
height, width = im.shape[:2]

Answer 4

You have an image of size wx h. Let's assume a roi size of 11 pixel per side, so offset = 5

Access the pixel from (w/2 - width/2, h/2 - height/2) to (w/2 + width/2, h/2 + height/2).

Then calculate the mean of all extracted pixel (so you're more robust to color variation).

You can of course change the color space if you want it in other space color depending of the kind of picture are you going to analyze.

Answer 5

This will find the RGB value of the pixel in the center of image using OpenCV.

import cv2
import numpy as np

img = cv2.imread("image.png")

height, width, depth = img.shape
circle_img = np.zeros((height, width), np.uint8)

mask = cv2.circle(circle_img, (int(width / 2), int(height / 2)), 1, 1, thickness=-1)
masked_img = cv2.bitwise_and(img, img, mask=circle_img)

circle_locations = mask == 1
bgr = img[circle_locations]

rgb = bgr[..., ::-1]

print(rgb)

Source