All right, let's take this a step at a time. You say " if i add matrix(same size of original image)as a noisy matrix". Ok, let's call this matrix "noiseImage" just to be descriptive.
Next you say " in original image,". OK that is confusing. Let me assume that you meant "to the original noise-free image." Let's call the original noise-free image "originalImage". So now you're saying let's do this operation: noiseImage+originalImage.
Then you say " then how could i get noise of image(original+noisy matrix) in matlab?" Well, first of all the noise of the image is the variable noiseImage, which we already had or else we couldn't do the addition operation. So you ALREADY HAVE it! But then you imply that the noise of (in) the image is originalImage+noiseImage, which is not true. That is your output image, not your noise. So now we have a situation with ambiguous and conflicting requests and leave us to guess or ask you for clarification.
If you have all three images (1) noiseImage, (2) originalImage, and (3) noisyImage or outputImage, then given any two, you can compute the third. Is that your situation? Do you have only two of these images, or even worse, only one of the images (presumable the noisy output image) and you want to compute the other two given only one image? If you have only the noisy image, there are a number of ways that you guess at what the noise might be - they all require that you make some assumptions about the noise.
Please clarify so we can help you further.
