calculating the average in a matrix

I want to calculate the average of a 3x3 matrix and replace the middle value with the calculated average of it and the 8 around it. I also want to repeat this in order to 'smooth' an image. my image is a 200x200 matrix read from a png file.( I read it using imread earlier). so what I am doing (or trying to do) is going through the whole image and averaging out the values in 3x3 matrices. (I am ignoring the border ones because they don't have 8 values surrounding them) This is the code that I wrote:

[rows, cols, colors]= size(image{1});
for rows=2:1:199, cols=2:1:199;
    ave=mean(mean([image{1}(rows,cols), image{1}(rows+1, cols), image{1}(rows-1, cols), image{1}(rows,cols+1), image{1}(rows,cols-1), image{1}(rows-1,cols+1), image{1}(rows-1,cols-1), image{1}(rows+1, cols+1), image{1}(rows+1, cols-1)]));
    image{1}(rows,cols,colors)= ave;
end

firstly, can anyone see something wrong with the code?

I then wanted to add up all of the pixels that were above 220 in this new image so I wrote :

      sum(sum(image{1}>220))

But I get 0 when im pretty sure im supposed to get a few. So there is something wrong somewhere with my coding and I think it has to do with the line

image{1}(rows,cols,colors)= ave;

or maybe my for loop is wrong.

Any help is greatly appreciated!

Thanks

Answer 1

[Sean de Wolski]

I = your_image;
Ismoothed = cast(conv2(double(I),ones(3)/9,'valid'),class(I));

A windowing average is just a convolution with a kernel of ones(size_window)/n_elements_in_window.

EDIT: changed same to valid since OP is ignoring border values.

[Jack Williams]

what you did is right but not exactly what I'm looking for, I have to do it with the averaging method, perhaps my use of the term 'smoothing' is misguided.

do you know what I am doing wrong here?

[Sean de Wolski]

No. I don't I didn't look at your code. Unless this is a homework assignment and you're forced to do this with double for-loops then there is no reason to do a windowing average in that manner.

[Sean de Wolski]

There still isn't any reason to use cell arrays even if you have to do double for-loops

for ii = size(I,1)-1:-1:2;

for jj = size(I,2)-1:-1:2

Ismoothed(ii,jj) = sum(sum(I(ii-1:ii+1,jj-1:jj+1)))/9;

end

end