Filling holes in image with OpenCV or Skimage

Question

I m trying to fill holes for a chessboard for stereo application. The chessboard is at micro scale thus it is complicated to avoid dust... as you can see :

Thus, the corners detection is impossible. I tried with SciPy's binary_fill_holes or similar approaches but i have a full black image, i dont understand.

Or invert the image, use find contours, fill them black and invert back again — PSchn, Sep 24 '16 at 16:06

Mazdak · Accepted Answer · 2018-06-27T07:04:54.130

4

Here is a function that replaces the color of each pixel with the color that majority of its neighbor pixels have.

import numpy as np
import cv2

def remove_noise(gray, num):
    Y, X = gray.shape
    nearest_neigbours = [[
        np.argmax(
            np.bincount(
                gray[max(i - num, 0):min(i + num, Y), max(j - num, 0):min(j + num, X)].ravel()))
        for j in range(X)] for i in range(Y)]
    result = np.array(nearest_neigbours, dtype=np.uint8)
    cv2.imwrite('result2.jpg', result)
    return result

Demo:

img = cv2.imread('mCOFl.png')
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

remove_noise(gray, 10)

Input image:

Out put:

Note: Since this function replace the color of corner pixels too, you can sue cv2.goodFeaturesToTrack function to find the corners and restrict the denoising for that pixels

corners = cv2.goodFeaturesToTrack(gray, 100, 0.01, 30)
corners = np.squeeze(np.int0(corners))

edited Jun 27 '18 at 07:04

answered Sep 24 '16 at 14:46

Mazdak

105,000
18
159
188

Thanks for your interesting results! :) It works well but the drawback is the diffusion at the corners conducting to an increase of errors i think. Perhaps i could erode after your correction? – user3601754 Sep 24 '16 at 14:49
I thought to a median filter but it conducts to the same problem. – user3601754 Sep 24 '16 at 14:51
1

@user3601754 Yes I was about to add an explanation regarding that. I think There is a way for refusing that by finding the corners and escaping the operation on them. – Mazdak Sep 24 '16 at 14:52
Yep i cant detect the corners with "cv2.findChessboardCorners" function :( – user3601754 Sep 24 '16 at 14:58
Ismt this nearly the same like just eroding the image? – PSchn Sep 24 '16 at 19:43
@PSchn Yeah, kind of. – Mazdak Sep 24 '16 at 21:35
So why not using opening instead?! The "holes" would be filled and the shape intact. Or use the method i mentioned above. – PSchn Sep 24 '16 at 21:54

score 2 · Answer 2 · answered Sep 25 '16 at 12:53

2

You can use morphology: dilate, and then erode with same kernel size.

answered Sep 25 '16 at 12:53

Andrey Smorodov

10,649
2
35
42

score 0 · Answer 3 · answered Oct 08 '22 at 16:32

0

A faster, more accurate way is to use skimage.morphology.remove_small_objects docs

im = imread('a.png',cv2.IMREAD_GRAYSCALE)
im = im ==255
from skimage import morphology
cleaned = morphology.remove_small_objects(im, 200)

answered Oct 08 '22 at 16:32

yazan sayed

777
7
24

Filling holes in image with OpenCV or Skimage

3 Answers3