What should my approach be towards designing a filter that can simultaneously remove salt & pepper noise and improve the resolution?

  • 4
    $\begingroup$ Your question needs some more details. "Resolution" is a somewhat vague term; do you mean you want to remove speckle noise while sharpening the image at the same time? $\endgroup$ – Jason R Dec 10 '11 at 17:31
  • $\begingroup$ yes that would be it $\endgroup$ – vini Dec 11 '11 at 9:27
  • 2
    $\begingroup$ Why does it need to be done simultaneously ? Why not use two filters ? $\endgroup$ – Paul R Dec 11 '11 at 12:27
  • $\begingroup$ am sorry for giving you such a hard time i am new to MATLAB and image processing.. Yes two or more filters that would get the job done is what i want to undertake $\endgroup$ – vini Dec 11 '11 at 15:51

If you use two linear filters, they could be combined (because of linearity, the convolution with the convolution of the two filters equals to the consecutive convolution with the two filters).


Nevertheless the simplest way to deal with salt & pepper noise is median filtering, which is non-linear. If you use linear filters it is possible.

| improve this answer | |
  • 2
    $\begingroup$ One issue with using linear filters is that removing salt and pepper noise is a lowpass operation, which will reduce image sharpness. A linear filter used for sharpening would be highpass, which would pass the undesired noise. I'm guessing a nonlinear approach would be required to effectively achieve both goals. $\endgroup$ – Jason R Dec 11 '11 at 14:30
  • $\begingroup$ Yes thanks a lot will keep your post in mind $\endgroup$ – vini Dec 11 '11 at 15:52
  • $\begingroup$ Which is the best filter by the way to reduce noise? $\endgroup$ – vini Dec 11 '11 at 15:54
  • 4
    $\begingroup$ That's, again, a very vague question. It's like asking "what is the best car?" It all depends upon the problem at hand. You might try posting some examples of what you'd like to process. $\endgroup$ – Jason R Dec 11 '11 at 22:11
  • $\begingroup$ There are several different types of noise. en.wikipedia.org/wiki/Image_noise As @jason-r said, you need to choose the best tool for the task at hand. Usually what people refer generically to as "noise" is Gaussian. $\endgroup$ – David Poole Dec 15 '11 at 13:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.