CODEWITHSUNDEEP
matlabimage-processing
user11459362
user11459362

Reputation:

How to change a cluster of 'on' pixels above a certain cluster size into 'off' pixels

I've made some MATLAB code that turns an image (stars) into a binary image by using a set threshold. It then finds and labels the clusters of 'on/1/white' pixels that are connected to each other and then gives an output like:

[1 1 1 0 0 0 0 0 0
 1 1 0 0 0 2 2 2 0
 0 0 0 3 3 0 2 0 0
 0 0 0 3 3 0 0 0 0]

To create the clusters the code uses a counter to give each cluster it's own unique ID, such as 1, 2 or 3 etc. However I now want to be able to turn the pixel clusters that are above a certain size, e.g. larger than 12 pixels in size, into 'off/0/black' pixels and remove them from the output.

Does anyone know how I would do this?

My code is shown below.

visited = false(size(binary_image));    % initialise an array with same size as image array that is logical. This records which pixels have been visited.
[rows, cols] = size(binary_image);
B = zeros(rows, cols);    % initialise an output array with all 0's that is the same size as the image array. Any 0's left don't belong to connected pixels.
ID_counter = 1;    % Labels connected pixels with unique ID's and keeps track of given ID's
for row = 1:rows    % search through rows of binary_image
    for col = 1:cols    % search through columns of binary_image
        if binary_image(row, col) == 0
            visited(row, col) = true;    % if position == 0 mark as visited and continue
        elseif visited(row, col)
            continue;    % if already visited position then continue
        else
            stack = [row col];    % if not visited before create stack with this location

            while ~isempty(stack)    % while stack isn't empty
                loc = stack(1,:);
                stack(1,:) = [];    % remove this location from stack

                if visited(loc(1),loc(2))
                    continue;    % if stack location already visited then continue
                end


                visited(loc(1),loc(2)) = true;    % is not visited before then mark as visited
                B(loc(1),loc(2)) = ID_counter;    % mark this location in output array using unique ID

                [locs_y, locs_x] = meshgrid(loc(2)-1:loc(2)+1, loc(1)-1:loc(1)+1);    % given this location, check 8 neighbouring pixels (N,E,S,W,NE,NW,SE,SW)
                locs_y = locs_y(:);
                locs_x = locs_x(:);

                out_of_bounds = locs_x < 1 | locs_x > rows | locs_y < 1 | locs_y > cols;    % get rid of locations that are out of bounds of image
                locs_y(out_of_bounds) = [];
                locs_x(out_of_bounds) = [];

                is_visited = visited(sub2ind([rows cols], locs_x, locs_y));    % get rid of locations already visited
                locs_y(is_visited) = [];
                locs_x(is_visited) = [];

                is_1 = binary_image(sub2ind([rows cols], locs_x, locs_y));    % get rid of locations that are 0
                locs_y(~is_1) = [];
                locs_x(~is_1) = [];

                stack = [stack; [locs_x locs_y]];    % add remaining locations to stack


            end

            ID_counter = ID_counter + 1;    % increase the unique ID by 1 after every cluster labelling
        end
    end
end

Upvotes: 1

Views: 88

Answers (1)

Wick
Wick

Reputation: 314

Just add a loop at the end where you examine how many pixels there are with each index.

pixel_limt = 12;
for jj = 1:length(ID_counter)
    id_index = find(B == jj);
    if numel(id_index) > pixel_limit
        B(id_index) = 0;
    end
end

Upvotes: 0

Related Questions