vectorization of nested for loop

i want to transfer an image pixels coordination into matrix (rows x 2) , this is the code with for loops, i want to vectorize it because it takes so much time for big images.

I=imread('1.bmp');

[rows,columns,~]=size(I);

pix_im=[];


for l=1:rows
    for c=1:columns
        pix_im=[pix_im;l c];
    end
end

Answer 1

[Ameer Hamza]

No wonder this code runs slow. You are using dynamic memory allocation. You can get several-fold speed gain just by pre-allocation: https://www.mathworks.com/help/matlab/matlab_prog/preallocating-arrays.html. However, you don't even need a loop for this. You can do it much more efficiently like this.

I=imread('1.bmp');
[rows,columns,~]=size(I);
[c,r] = ndgrid(1:columns, 1:rows);
pix_im = [r(:) c(:)]

[chabani mohammed ali]

thank you so much it works and it's fast just like i want.

i was trying to do the same thing with meshgrid but i couldn't get a correct result ( am new to matlabe),

and even when i use a loop with pre-allocation the way you proposed is faster. thank you again.

[Ameer Hamza]

I am glad to be of help!

[chabani mohammed ali]

i posted another question if you don't mind to take a look on it https://fr.mathworks.com/matlabcentral/answers/648293-pre-allocation-of-array-used-like-a-fifo-stack/?s_tid=ans_lp_feed_leaf