circle-finder.py

import cv2
import numpy as np

# Load image, grayscale, median blur, Otsus threshold
image = cv2.imread('film-10176.tif')
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
blur = cv2.medianBlur(gray, 11)
thresh = cv2.threshold(blur, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)[1]

# Morph open
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5,5))
opening = cv2.morphologyEx(thresh, cv2.MORPH_OPEN, kernel, iterations=3)

font = cv2.FONT_HERSHEY_SIMPLEX

# org
org = (500, 50)

# fontScale
fontScale = 1

# Blue color in BGR
color = (255, 0, 0)

# Line thickness of 2 px
thickness = 2

# Using cv2.putText() method





# Find contours and filter using contour area and aspect ratio
cnts = cv2.findContours(opening, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
cnts = cnts[0] if len(cnts) == 2 else cnts[1]
for c in cnts:
    peri = cv2.arcLength(c, True)
    approx = cv2.approxPolyDP(c, 0.04 * peri, True)
    area = cv2.contourArea(c)
    if len(approx) > 5 and area > 1000 and area < 500000:
        ((x, y), r) = cv2.minEnclosingCircle(c)
        cv2.circle(image, (int(x), int(y)), int(r), (36, 255, 12), 2)

string = "Radius is: " + str(int(r))+"px"
print("Radius is: " + str(int(r))+"px")
image = cv2.putText(image, string, org, font,
                    fontScale, color, thickness, cv2.LINE_AA)


cv2.imshow('thresh', thresh)
cv2.imshow('opening', opening)
cv2.imshow('image', image)
cv2.waitKey()