Determining the number of squares with an intersection?

Question

Python

You need to find the number of squares with intersection in the picture.
Now there is this code:

import cv2
import numpy as np

img = cv2.imread('./task1.bmp', cv2.IMREAD_UNCHANGED)

#convert img to grey
img_grey = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
#set a thresh
thresh = 100
#get threshold image
ret,thresh_img = cv2.threshold(img_grey, thresh, 255, cv2.THRESH_BINARY)
#find contours
contours, hierarchy = cv2.findContours(thresh_img, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
# print(len(contours))
#create an empty image for contours
img_contours = np.zeros(img.shape)
# draw the contours on the empty image
cv2.drawContours(img_contours, contours, -1, (0,255,0), 3)
#save image
cv2.imwrite('./res.bmp',img_contours)

Original image:


Result after processing


As you can see, the squares in the pictures intersect and the program finds their common contours. It is necessary that the program find exactly the contours of the squares and count them.

Answer 1

Do you know the size of the square? Use a sliding window.
Cut out in turn all fragments of the image of this size, if it is completely black, then we fix the square in this position.
There may be false positives if two overlapping squares are perfectly aligned on the same vertical/horizontal. Here you need to search for all triggers in this area, and select the extreme left / right or bottom / top.

Answer 2

Everything is done literally in three teams.
1. Closing with cv2.morphologyEx, the core is square, 2 times smaller than squares. Single squares are removed.
2. inversion. cv2.bitwise_not
3. cv2.connectedComponents count remaining connected squares.
If you need to select individual squares in the union, then use cv2 after step 1. matchTemplate