bonevale's profile - activity

So far I am thinking training cascade. If anyone has other ideas please comment, thank you.

I had a pretty tough time installing opencv the first time around so I am not planning on rebuilding it, however thanks for the suggestion. I am looking more so for pseudo code or an algorithm which I can implement myself if need be.

I am looking to count the number of little circles or ovals on the lines. I am relatively new to opencv so I am open to any suggestions that get me started thank you. This is the image. C:\fakepath\ROOT003 - 1200 DPI.jpg

Currently I am implementing the zhang-suen method but my images are 300 ppi and it takes a couple of minutes. Can someone point me in the right direction. I need the skeleton to be connected and there are parallel lines so the the algorithm I've seen in python that uses open and closing is not good. Thanks

Yes when I call rectangle to draw around the contours instead of drawContours I get the behavior I want where all the little dots are selected so the rest of my function works as desired. And yes before all this my image is given a threshold.

I found the post and tried the tip but it didn't fix my error, thanks though.

here's my code

gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
labelNum, labelImg, contours, GoCs = cv2.connectedComponentsWithStats(gray)

    for c in contours:
        x,y,w,h,size = c
        if size <= 100:
            cv2.drawContours(image, [c], 0, 0)