Removing noise of an image

I want to remove noise using the following example

http://nl.mathworks.com/help/images/noise-removal.html

but I receive error message after:

I = imread('eight.tif');
>> figure

imshow(I)

Warning: Image is too big to fit on screen; displaying at 50% 
> In images.internal.initSize (line 71)
  In imshow (line 305) 
>> J = imnoise(I,'salt & pepper',0.02);
figure
imshow(J)
Warning: Image is too big to fit on screen; displaying at 50% 
> In images.internal.initSize (line 71)
  In imshow (line 305) 
>> Kaverage = filter2(fspecial('average',3),J)/255;
figure
imshow(Kaverage)

Undefined function 'conv2' for input arguments of type 'double' and attributes 'full 3d real'.

Error in filter2 (line 59)
        y = conv2(hcol, hrow, x, shape);

What do I have to correct?

Answer 1

[Walter Roberson]

You cannot apply conv2() to all of an RGB image at once. You need to process it one channel at a time.

[YaYaYa87]

How to process it one channel at a time? How to correct the script?

[Walter Roberson]

for chan = 1 : size(J,3)
  Kaverage(:,:,chan) = filter2( fspecial('average',3), J(:,:,chan)) /255;
end

[YaYaYa87]

Can you show me the whole script,please,because I see error message again?

[YaYaYa87]

don't need to write it. I've managed to fix it. I just converted the image into grayscale. Anyway, thank you for the answers!

Answer 2

[Image Analyst]

This code works perfectly fine:

I = imread('eight.tif');
imshow(I)
[rows, columns, numberOfColorChannels] = size(I)
J = imnoise(I,'salt & pepper',0.02);
figure
imshow(J)
Kaverage = filter2(fspecial('average',3),J)/255;
figure
imshow(Kaverage)

All I added was a call to size() to show how many color channels it is. It shows

rows =
   242
columns =
   308
numberOfColorChannels =
     1

Note that numberOfColorChannels = 1 which means it's grayscale. Yours must be 3 which means it's color. Because your J, and therefore I, are color, this proves that somehow either

1. you sometime prior overwrote your version of the standard demo image 'eight.tif' with a color version of it,
2. or there is code in your script that you did not show us.

What I would do is to reinstall MATLAB to restore your original, proper demo images.

[YaYaYa87]

Yes, mine image is coloured and i turned it into grayscale: RGB = imread('eight.tif'); >> imshow(RGB) >> I = rgb2gray(RGB);

That's it :)

[Image Analyst]

Again, reinstall. You should not want corrupted/changed versions of the standard demo images. As you've seen, it has already led to problems and will lead to problems again in the future if you don't restore them to their original version.