How to find and store each spot center position in the loop.

Hello, I would like to seek your help. I have several image signal, and would like to find and stored each spot center position in the loop. I tried to write a code as below, but I got the error message "

s =

3×1 struct array with fields:

Centroid

Expected one output from a curly brace or dot indexing expression, but there

were 3 results.

My code as below,

close all; clear all; clc
%%---------Open image-------
D = 'C:\Users\Admin\Desktop';
S = dir(fullfile(D,'*.bmp')); % pattern to match filenames.

CCDx = 1:numel(S)-1;
CCDy = 1:numel(S)-1;
 
for k = 1:numel(S)-1
     F = fullfile(D,S(k).name);
     I = imread(F);
%      figure(),imshow(I);
     
%%-------open image BG file----
    S_BG = dir(fullfile(D,'BG1.bmp')); % pattern to match filenames.
    for k_BG = 1:numel(S_BG)
        F_BG = fullfile(D,S_BG(k_BG).name);
        I_BG = imread(F_BG);
%          figure(100),imshow(I_BG)
        
%%%---convert data type from RGB to gray-----
         Z = im2double(I);
         Z = rgb2gray(Z);
        Z_BG = im2double(I_BG); 
        Z_BG = rgb2gray(Z_BG);
%           figure(200),imshow(Z_BG)

%%%------cut background-------
        Z_net = Z-Z_BG;
        
%%%-----croppedImage-------- 
        leftColumn=115; % x axis
        width = 100;
        topLine = 180; % y axis
        height = 100;
        croppedImage = imcrop(Z_net, [leftColumn, topLine, width, height]);
        
        bw = imbinarize(croppedImage);
        s = regionprops(bw,'Centroid')        

%%%-------stored and save position for each image----------        
        px = s.Centroid(1);
        py = s.Centroid(2);
 
        figure()
        imshow(croppedImage,'InitialMagnification','fit')
        hold on
        plot(s.Centroid(1),s.Centroid(2),'b*')
 
        CCDx(k) = px;
        CCDy(k) = py;
    end     
end

Answer 1

[Walter Roberson]

Use

        bw = bwareafilt(imbinarize(croppedImage), [10 inf]);

Some of your images end up with very small extra regions, just a few pixels each.

[Tina Hsiao]

I got the error merrage shows Expected one output from a curly brace or dot indexing expression, but there were 2 results.

How can I solve it?

The code as below,

close all; clear all; clc

%%---------find peak program-------
D = 'C:\Users\Admin\Desktop';
S = dir(fullfile(D,'*.bmp')); % pattern to match filenames.

CCDx = 1:numel(S)-1;
CCDy = 1:numel(S)-1;
 
for k = 1:numel(S)-1
     F = fullfile(D,S(k).name);
     I = imread(F);
%      figure(1),imshow(I)
     
%%-------open image BG file----
    S_BG = dir(fullfile(D,'BG1.bmp')); % pattern to match filenames.
    for k_BG = 1:numel(S_BG)
        F_BG = fullfile(D,S_BG(k_BG).name);
        I_BG = imread(F_BG);
%          figure(100),imshow(I_BG)
        
%%%---convert data type from RGB to gray-----
         Z = im2double(I);
         Z = rgb2gray(Z);
        Z_BG = im2double(I_BG); 
        Z_BG = rgb2gray(Z_BG);
%           figure(200),imshow(Z_BG)

%%%------cut background-------
        Z_net = Z-Z_BG;
        
%%%-----croppedImage-------- 
        leftColumn=115; % x axis
        width = 100;
        topLine = 180; % y axis
        height = 100;
        croppedImage = imcrop(Z_net, [leftColumn, topLine, width, height]);
        
        bw = bwareafilt(imbinarize(croppedImage), [10 inf]);
        s = regionprops(bw,'Centroid');        

%%%-------stored and save position for each image----------        
        px = s.Centroid(1);
        py = s.Centroid(2);
 
        figure()
        imshow(croppedImage,'InitialMagnification','fit')
        hold on
        plot(s.Centroid(1),s.Centroid(2),'b*')
 
        CCDx(k) = px;
        CCDy(k) = py;
    end     
end

[Walter Roberson]

One of your images (not one you attached above) has an extra blob that is 10 or more pixels.

If you change to

        bw = bwareafilt(imbinarize(croppedImage), 1);

then it will extract only the largest blob, but I am concerned about the possibility that in time one of your images might have two non-trivial blobs that you need to take into account. I think you should put in a conditional breakpoint on the assignment to px, to stop if numel(s) > 1, at which point you should examine the image stored in bw and determine whether the extra blobs are significant or not. You could consider raising the [10 inf] to something larger to filter out more small blobs, but you should be having a closer look at the data before you determine whether it is wise.

You could consider using imclose() with a structuring element in order to join nearby blobs together. But I still worry that at some point there might be multiple significant blobs, and that you should plan your code for that possibility.

[Tina Hsiao]

Thanks for suggestion and help. It works!