How do you apply a 3x3 median filter on an image that is noisy (salt and pepper) , and what is it supposed to do?

42 views (last 30 days)
is this correct?
inpict = imread('Q1_noisy_image (1).tif');
A = medfilt2(inpict,[3 3]);
imshow(A)

Answers (2)

Image Analyst
Image Analyst on 2 Dec 2023
If inpict is a gray scale image, that would apply a median filter to ALL pixels in the image, not just the impulse/salt&pepper noise.
A median filter's claim to fame is that it smooths noisy areas but leaves true edges sharp (it doesn't blur step edges).
If you want a median filter to operate on only salt and pepper noise, you have to
  1. First median filter the whole image
  2. Get a mask of the salt and papper noise, where values are either 0 or 255 or whatever values you want to use.
  3. Replace the original image with the median filtered values but ONLY within the noise mask, not everywhere.
Something like
filteredImage = medfilt2(grayImage, [3,3]);
mask = (grayImage == 0) | (grayImage == 255);
grayImage(mask) = filteredImage(mask);
For a full demo, see my attached demos. For a color image, do it one color channel at a time.

DGM
DGM on 2 Dec 2023
Edited: DGM on 2 Dec 2023
  2 Comments
DGM
DGM on 2 Dec 2023
Edited: DGM on 2 Dec 2023
In the future, try to keep things about one question in a common thread so that efforts don't get divided. It's okay to revise your question either by editing it or by using comments.
FWIW, the link I'd provided in 2054999 includes a link to the demos that Image Analyst also posted above.

Sign in to comment.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!