The distance from center of the binary image to the furthest point of the contour

The main idea is to find the furthest point coordinates of the contour from the center of the image in Python.

Normally it is possible to find the extreme points of the contour using findcontours and argmin and argmax in OpenCV.

However, for this problem we need the most distant point from the center of the image as illustrated in the image below:

.

How can we get this distance?

Seems straightforward enough:

import cv2 as cv
import numpy as np

file = r'/Users/lh/Downloads/LTSmL.png'
img = cv.imread(file, 0)

max_intensity = np.max(img)

mask = np.argwhere(img == max_intensity)
center = (img.shape[0]/2, img.shape[1]/2)

maxdist = 0
maxpoint = None

for point in mask:
    dist = np.sqrt((center[0] - point[0])**2 + (center[1] - point[1])**2)
    if dist > maxdist:
        maxdist = dist
        maxpoint = point

img_color = cv.cvtColor(img, cv.COLOR_GRAY2RGB)

img_color[maxpoint[0] - 5:maxpoint[0] + 5, maxpoint[1] - 5:maxpoint[1] + 5] = [0, 0, 255]

cv.imshow('Farthest point in red', img_color)
cv.waitKey()
cv.destroyAllWindows()

print(dist, point)

You start by masking the image to find the pixels making up the circle, then you calculate the distance to each of these points and return the max distance as well as the corresponding point.

Edit: You can probably do this way more elegantly using np.linalg.norm and np.argwhere, but the approach stays the same.