How to use CV2 to cut portions of an image into separate images?

user_6067928 0
Viewed 178

enter image description hereI have an image which contains many same sized rectangular panels separated by whitespace. I want to separate each rectangular panel into its own image. Any suggestions on how to go about it?

I thought of using cv2.floodfill to fill in the whitespace between panels and then invert the mask. However, that creates just a single mask and not an array of masks which I could use to identify coordinates and then split the panels into images.

Any pointers on how to do this smartly would be much appreciated.

opencv image-processing python
1 Answers

You can try applying contours on an image, Use binary image cv2.THRESH_BINARY than Gray image for effective contours -> apply contours

Filter contours

1st Filter - which are first level with no parent Contours Hierarchy,

2nd Filter - which are with same range of rectangles with 1% deviation

Followed by Bounding box of contours and slicing image,

import numpy as np
import cv2 as cv
filename = r'UcbZn.jpg'
im = cv.imread(filename)
imgray = cv.cvtColor(im, cv.COLOR_BGR2GRAY)
ret, thresh = cv.threshold(imgray, 240, 255, cv.THRESH_BINARY)
contours, hierarchy = cv.findContours(thresh, cv.RETR_TREE, cv.CHAIN_APPROX_SIMPLE)

# Filter -1 Filtering countours with first level with no parent
a = [[cv.contourArea(c), h, c] for c, h in zip(contours, hierarchy[0]) if h[3]==0]
a.sort(key=lambda x: -x[0])

base_contour = a[0]
# Filter-2 contours of rectangle size
res = [c for c in a if c[0] >= base_contour[0]*0.99]


count = 0
for c in res:
    x,y,w,h = cv.boundingRect(c[2])
    cv.imwrite('objects_{}.jpg'.format(str(count)), im[y:y+h,x:x+w, :])
    count+=1

enter image description here

user_82080060
Related