How to count grain in which the image is eroded?

Hi, I'm a student and doing a project in image processing. I have to label and count this image .

Answer 1

[Walter Roberson]

Unless you expect the count to be something other than 12, you can use bwlabel()

[Jothika Charlees]

The grains are correctly 12 then what to do?

[Walter Roberson]

bwlabel() with two outputs returns both the labeled image and the number of regions.

[Jothika Charlees]

clc;
clear all;
close all;
a=imread('grain 2.jpg');
figure
imshow(a);
title('Input image');
b=rgb2gray(a);
figure
imshow(b);
title('Grayscale image');
c=medfilt2(b,[3 3]);
figure
imshow(c);
title('Median filtered image');
d = im2bw(b,0.5);
figure
imshow(d);
title('Binarized image');
se = strel('disk',11);        
e = imerode(d,se);
figure
imshow(e);
title('Eroded image');
[labeledImage, numberOfBlobs] = bwlabel(e, 8);
coloredLabelsImage = label2rgb (labeledImage, 'hsv', 'k', 'shuffle'); 
figure;
imshow(coloredLabelsImage);
title('Labeled Image');

this is the code i used but i don't know how to label the image grain with numbers. The attachment shows the output image that i get using this code

[Walter Roberson]

Yes, that is a labeled image in MATLAB terms. Each different area has been given a different number, and each different number has been represented as a different color based upon the current colormap() .

Perhaps what you wanted was some text written onto each blob. If so then use regionprops() to request the Centroid . Loop through the returned information, pull out the Centroid information, and text() the appropriate label number at that location.