How to detect the nearest square to the image center points?

Question

How do I discover the nearest square to the center of the image?

I already have the square's vertices and the center, just need to find out which square is closest to the center.

I have a problem similar to this image. I would like to select the nearest square (the red square).

Answer 1

This is the solution I have for my problem. Where the x0 and y0 can be the coordinates of the center of my image and the x1 and y1 will be center of coordinates of the square center.

import math
def dist(x0, y0, x1, y1):
    a = (x1 - x0)**2 + (y1 - y0)**2
    b = math.sqrt(a)
    return b
print dist(5, 5, 4, 6)
print dist(5, 5, 9, 2)

Answer 2

Two possibilities, of which one is mentioned in your answer and the comment by @miki

1. Square center

You do not need to implement your own distance function. SciPy already has a few. For example your euclidean distance in scipy.spatial.distance.euclidean:

from scipy.spatial import distance
distance.euclidean((x0, y0), (x1, y1))

No need to reinvent the wheel.

2. Square edges

In the following example it can be argued whether the red or the blue square are closer to the center. By euclidean distance, it is the blue one. The red one is overlapping the center, though.

If you wanted to have the square with the closest pixel to the center, you could do something like

square = (upper_left_x, upper_left_y, lower_right_x, lower_right_y)
center = (x, y)

if upper_left_x <= x <= lower_right_x and upper_left_y <= y <= lower_right_y:
    return 0 # point in square
elif upper_left_x <= x <= lower_right_x: # on vertical line
    return min(upper_left_y -y, lower_right_y - y)
elif upper_left_y <= y <= lower_right_y: # on horizontal line
    return min(upper_left_x -x, lower_right_x - x)
else:
    points = []
    for x in (upper_left_x, lower_right_x):
        for y in (upper_left_y, lower_right_y):
            points.append((x,y))
    return min([distance.euclidean((x,y), p) for p in points])

Original answer before edit to question

You can split this up:

1. import image
2. find the squares
   1. find the corners
   2. connect them
3. compute distance to image center

The main point here may be any of them (except for 1.2 probably).

There is a neat OpenCV Python tutorial, which tells you how to do some of that. Let us start:

import

import cv2
img = cv2.imread('L3h9H.png')

imports the image.

To see that you imported correctly, you can use

from matplotlib import pyplot as plt
plt.imshow(img, cmap = 'gray', interpolation = 'bicubic')
plt.xticks([]), plt.yticks([])  # to hide tick values on X and Y axis
plt.show()

which shows it to you.

find corners

import numpy as np
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
gray = np.float32(gray)
dst = cv2.cornerHarris(gray, 2,3,0.04)
for i, valx in enumerate(dst):
    for j, valy in enumerate(valx):
        if valy > 0:
            print '%s, %s: %s' % (i, j, valy)

Finds corner using one of the built-in algorithms. The corners are listed afterwards.

Next steps would be:

• compute lines between corners (maybe).
• Show min distance of line to image center.

Say if you need more help.