How to find dots with OpenCV in Python?

user_676192 0
Viewed 289

I'm just starting with OpenCV so this question is more like "what do you call the thing in OpenCV that does..." than "how do you do the thing..." - if I know what to look for usually docs and more googling go a long way.

So: using OpenCV in Python how could I find the position of all the black dots at the border between red and gray?

It's a nice to have if it is possible to tell if the border is horizontal or vertical or alternatively its angle at the point.

enter image description here

In other words - the position of the dots inside the white borders, but not the ones in the greyed out area:

enter image description here

opencv python
1 Answers

Some setup:

im = cv.imread("4V56c.png")
hsv = cv.cvtColor(im, cv.COLOR_BGR2HSV)

Dots:

# axis=2 is the color dimension. all values must be at most 30.
# result is a boolean array (0 or 1). times 255 because of reasons.
dots = (im <= 30).all(axis=2) * np.uint8(255)

dots

Border:

# a certain saturation
mask = (hsv[:,:,1] >= 50) * np.uint8(255)
# erase all the gray stuff
mask = cv.morphologyEx(src=mask, op=cv.MORPH_CLOSE, kernel=None, iterations=5)
# blow up and shrink
dilated = cv.dilate(mask, kernel=None, iterations=5)
eroded = cv.erode(mask, kernel=None, iterations=5)
# difference
border = dilated & ~eroded

border

Combined:

dots & border

dots & border

user_26028770
Related