Size filtering and show the position of centroids on the image

I have a binary image, BW and a grayscale image, I. I want to select only regions in the binary image with the same value that have an area > 50. For example there are 70 regions with area>50 and BW==1. I wonder how to show the position of the centroids of these regions with their number (1:70) on the image. Any suggestions?

Answer 1

[Image Analyst]

Size filtering and finding and displaying centroids is done in my demo "BlobsDemo" as well as some other useful things.

http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862

Otherwise, here's a shortened demo:

% Demo by Image Analyst, December, 2021.
clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clearvars;
workspace;  % Make sure the workspace panel is showing.
format long g;
format compact;
fontSize = 16;
fprintf('Beginning to run %s.m ...\n', mfilename);
numSpaces = 6;
myBoard = zeros(numSpaces+1,numSpaces+1);
computersBoard = zeros(numSpaces+1,numSpaces+1);
% Read in sample image
grayImage = imread('kobi.png');
if ndims(grayImage) >= 3
    grayImage = rgb2gray(grayImage);
end
% Resize this image to get our blobs in the range where 50 is a good blob area.
grayImage = imresize(grayImage, 0.6);
subplot(2, 2, 1);
imshow(grayImage, []);
impixelinfo
title('Original Image', 'FontSize', fontSize);
% Get histogram
subplot(2, 2, 2);
imhist(grayImage);
grid on;
title('Gray Level Histogram', 'FontSize', fontSize);
% Get mask
threshold = 65;
xline(threshold, 'Color', 'r', 'LineWidth', 2)
mask = grayImage < threshold;
% Fill holes
mask = imfill(mask, 'holes');
% Erode to separate them.
mask = imerode(mask, true(11));
subplot(2, 2, 3);
imshow(mask);
% Get areas of this mask
props = regionprops(mask, 'Area', 'Centroid');
allAreas = sort([props.Area])
centroids = vertcat(props.Centroid);
% Make title
caption = sprintf('%d blobs ranging from %d to %d pixels', length(props), min(allAreas), max(allAreas))
title(caption, 'FontSize', fontSize);
% Place crosshairs on the centroids.
xc = centroids(:, 1);
yc = centroids(:, 2);
hold on
plot(xc, yc, 'r+', 'MarkerSize', 15, 'LineWidth', 2);
% Now extract only those blobs with area >= 50 pixels using bwareaopen() or bwareafilt().
mask = bwareaopen(mask, 50);
subplot(2, 2, 4);
imshow(mask);
% Now redo the analysis.
% Get areas of this mask
props = regionprops(mask, 'Area', 'Centroid');
allAreas = sort([props.Area])
% Make title
caption = sprintf('%d blobs ranging from %d to %d pixels', length(props), min(allAreas), max(allAreas))
title(caption, 'FontSize', fontSize);
% Mark centroids on the image
for k = 1 : length(props)
    xc = props(k).Centroid(1);
    yc = props(k).Centroid(2);
    caption = sprintf('+ #%d at (%.1f, %.1f)', k, xc, yc);
    text(xc, yc, caption, 'Color', 'r', 'FontWeight', 'bold')
end
g = gcf;
g.WindowState = 'maximized'

Answer 2

[bym]

from the documentation:

I = imread('coins.png');
figure, imshow(I)

bw = im2bw(I, graythresh(getimage));
figure, imshow(bw)

bw2 = imfill(bw,'holes');
L = bwlabel(bw2);
s = regionprops(L, 'centroid');
centroids = cat(1, s.Centroid);

%Display original image and superimpose centroids.

imshow(I)
hold(imgca,'on')
plot(imgca,centroids(:,1), centroids(:,2), 'r*')
hold(imgca,'off')

[Hassan]

thanks proecsm for the comment. I want to show the number of region beside the centroid on the image as well. As I want to show the mean value of each region in a plot but want to show the position of the regions on the image.

[Image Analyst]

Obviously then you didn't look up my BlobsDemo. Any reason why not?

[Hassan]

thanks Image Analyst. your code was quite useful.

Answer 3

[Walter Roberson]

If centroid #K is at position x, y, then to label that point on the graph with that number, use

text(x,y,num2str(K));

[Hassan]

thanks Walter. this was what I wanted to complete proecsm's code for my work.

Answer 4

[Devie Nur AIni]

Ada yang tau ga kenapa pas coba aku running dia eror dan tulisannya

"check for missing argument or incorrect argument data tupe in call to function 'centroid'

% Peroleh pusat massa dan letakkan di tengah citra 
[xc, yc] = centroid(m,n); 
xc = round(xc); 
yc = round(yc); 
xc = xc - round((n/2)); 
yc = yc - round((m/2));

[Image Analyst]

@Devie Nur AIni, you have evidently, according to the error message, created a function called centroid, and that function does not expect two inputs m and n. Place your cursor in there and type control-D to edit that function. Or else do this on the command line:

>> edit centroid.m

I don't know what that function is supposed to do. Did you write it, or did someone else write it and you just put it on your computer?