0

I figured out how to detect the orientation of a semicircle. When my program gets the contours, I want it to retrieve them from left to right, however, OpenCV recognizes the semicircles on the bottom as a separate row, and counts the semicircles on the top first then the semicircles on the bottom (shown in the picture).

EDIT

lower = np.array([0, 0, 0])
upper = np.array([15, 15, 15])
shapeMask = cv2.inRange(img, lower, upper)

_, contours, _ = cv2.findContours(shapeMask.copy(), cv2.RETR_EXTERNAL,    cv2.CHAIN_APPROX_SIMPLE)

contours.reverse()

for cnt in contours:
    if len(approx)== 4 or len(approx) == 8:
          if(np.array_equal(img[midpoint[1],midpoint[0]],[0, 0, 0])):
              #unshaded square
          elif(np.array_equal(img[midpoint[1],midpoint[0]],[0, 255, 255])):
              #shaded square
      elif len(approx) == 3 or (len(approx) >= 14 and len(approx) <= 17):
          if(np.array_equal(img[midpoint[1],midpoint[0]],[0, 0, 0])):
              #unshaded triangle
          elif(np.array_equal(img[midpoint[1],midpoint[0]],[0, 255, 255])):
              #shaded triangle
      elif len(approx) >= 9 and len(approx) <= 13:
          #semicircle
      elif len(approx) > 18:
          if(np.array_equal(img[midpoint[1],midpoint[0]],[0, 0, 0])):
              #shaded circle
          elif(np.array_equal(img[midpoint[1],midpoint[0]],[0, 255, 255])):
              #unshaded circle

Here is the image I want to be able to differentiate between the semicircle on the top and the semicircle on the bottom, and count the bottom first as it is before the one on the top

Aneesh Kotnana
  • 61
  • 7
  • 1
    Can you add the code you use to create the circle and the contour code you are using? – lhcgeneva Dec 05 '15 at 16:48
  • have a look at http://stackoverflow.com/questions/20698613/detect-semi-circle-in-opencv/20706100#20706100 on how to analyze the semi-circle contour. It's C++ but maybe you can translate it. – Micka Dec 05 '15 at 18:56

0 Answers0