Segment grid boxes

rgb = imread('peppers.png');
imshow(rgb)
hold on

M = size(rgb,1);
  N = size(rgb,2);

    for k = 1:25:M
        x = [1 N];
        y = [k k];
        plot(x,y,'Color','w','LineStyle','-');
        plot(x,y,'Color','k','LineStyle',':');
    end

    for k = 1:25:N
        x = [k k];
        y = [1 M];
        plot(x,y,'Color','w','LineStyle','-');
        plot(x,y,'Color','k','LineStyle',':');
    end

    hold off

Suppose I want to segment and extract individual areas of the grid .. How do I go about it ? Like, which boundary detection algorithm is used for this purpose ?

[Sean de Wolski]

Post an image! How do you expect us to figure anything out when we can't see what you're doing!

[Siddharth Mallya]

http://img703.imageshack.us/i/gridg.png/

This is my image. If I want to segment each block of the grid, should I crop it each time ? or should I use some other approach ?

Answer 1

[Walter Roberson]

Regular rectangular grids, or potentially irregular spacing?

If it is a regular rectangular grid, crop to contain only full cells, detect the grid line separation, then mat2cell() or blkproc() or blockproc()

[Siddharth Mallya]

Right now I am using this in my program,

for i = fr1:31:lc1

for j = fc1:21:N

letter = imcrop (img, [j i 21 31]);

Works fine .. but takes quite a bit of time to process.

[Walter Roberson]

Your variables hint at "first row" to "last column" which doesn't sound right. A renaming might be in order.

Now, provided that lc1 and N have been chosen to be on grid boundaries:

mat2cell(img(fc1:N+21,fr1:lc1+31), 21 * ones(1,1+(N-fc1)/21), 31 * ones(1,1+(lc1-fr1)/31))

[Siddharth Mallya]

Thank you so much Walter ! You've been a great help to me .. :)

Answer 2

[Pedro Serra]

An answer to this question would be really nice. I'm dealing with the same issue:

I have an image upon which I've drawn a grid. I now would like to understand how can I extract individual areas of the grid?

(My final goal is to color specific areas of that grid)

Thanks in advance!

[Siddharth Mallya]

Well, one way I am thinking now is to crop the image in a for loop .. But i think the computational complexity increasing in this .. So ya, are there any boundary detection algorithm like flood fill with a seed point ?