Can anyone explain about "cv2.HoughCircles" method parameter in Simple language?

Python
1

I’m using OpenCV python library to find circles from concentric circle and successfully found all the circles but not able to understand the parameter of the “cv2.HoughCircles” method. I also hardcoded the range for “MinRadius” and “MaxRadius”. I just want to define the range for the radius based on the given image. I also go through the official document for “cv2.HoughCircles”
here
but this is also not understandale for me. Any help would be much appreciated!

Here is an image to find the circle:

GoldenSpike
GoldenSpike1563×661 142 KB

Here is my python code:

import numpy as np
import cv2
image = cv2.imread("GoldenSpike.png",0)
output = cv2.imread("GoldenSpike.png",1)
cv2.imshow("Original image", image)
cv2.waitKey()

blurred = cv2.GaussianBlur(image,(11,11),0)

cv2.imshow("Blurred image", blurred)
cv2.waitKey()
previous=0;
i=4
for maxR in range(9,425,9):
    # Finds circles in a grayscale image using the Hough transform
    circles = cv2.HoughCircles(blurred, cv2.HOUGH_GRADIENT, 1, 100,
                             param1=100,param2=90,minRadius=i,maxRadius=maxR)
    i+=4
    # cv2.HoughCircles function has a lot of parameters, so you can find more about it in documentation
    # or you can use cv2.HoughCircles? in jupyter nootebook to get that 

    # Check to see if there is any detection
    if circles is not None:
        # If there are some detections, convert radius and x,y(center) coordinates to integer
        circles = np.round(circles[0, :]).astype("int")

        for (x, y, r) in circles:
            # Draw the circle in the output image
            cv2.circle(output, (x, y), r, (0,255,0), 1)
            # Draw a rectangle(center) in the output image
            cv2.rectangle(output, (x - 2, y - 2), (x + 2, y + 2), (0,255,0), -1)

cv2.imshow("Detections",output)
cv2.imwrite("CirclesDetection.jpg",output)
cv2.waitKey()
2

According to the documentation, the method parameter has to be HOUGH_GRADIENT.

So HoughCircles detects the contours using the Canny operator, and for each point p and radius r it computes the sum of the contour pixels for a given circle pc:

Circ_Hough(p,r)=SUM(Canny(pc)==1 where d(p,pc=r))
# p - central pixel of a circle, r - radius, pc - pixels on the circle

Now, the param1 parameter is the canny threshold maxval. It means that the gradient value higher than param1 is always considered a contour pixel. For more detailed explanation see this tutorial.
You can check the effect of this parameter by running only the canny operator on the image:

cny = cv2.canny(image,param1/2,param1)
cv2.imshow("canny",cny)

The param2is the circle detection threshold. It means that for a point to be considered as a circle center, Circ_Hough(p,r)>param2. If param2 is too low, you’ll have false detections, if it’s too high, some good detections will be discarded.