Detecting multiple object in an image

Hello,

I am working on a project where I have to detect photovoltaic modules using infrared images.

I am using a template image to detect the modules but I can only detect one. What should I do to detect all the modules?

I do not have a lot of experience in programming but I am really trying.

[Image Analyst]

How are you detecting the one? Are you thresholding and calling bwlabel() to do the counting? Attach your image and your "single detection" code.

[Felipe Souza]

Original Image

[Felipe Souza]

Template

[Felipe Souza]

clear all; close all; clc;

%% READING THE RGB IMAGE
A = imread('Teste.jpg');
imshow(A);

%% CONVERTING RGB IMAGE TO GRAYSCALE FORMAT
A_gray = rgb2gray(A);
imshow(A_gray);

%% CONVERTING IMAGE TO DOUBLE FORMAT
A_gray = im2double(A_gray);
imshow(A_gray);

%% CONVERTING THE GRAYSCALE IMAGE TO BINARY FORMAT
otsu_level = graythresh(A_gray);
A_otsu_thresh = im2bw(A_gray, otsu_level);
imshow(A_otsu_thresh);

%% DETECTING OBJECTS IN THE IMAGE
label=bwlabel(A_otsu_thresh);
max(max(label))
for j=1:max(max(label))
  im1 = (label==j);
  figure(2), subplot(6,6,j);imshow(im1); title(j);
  j= j+1;
end

%% CORRELATION
[Ir Ic]=size(A_gray);
T1 = imread('Template.jpg'); %read template of the image
T = rgb2gray(T1);
imshow(T);
[Tr Tc]=size(T);
R=normxcorr2(T,A_gray); %find normalize cross correlation
R=imcrop(R,[Tc Tr Ic Ir]); %crop extra pixels
[row column value] = find(R==(max(max(R)))) %find the coordinates where maximum correlation exists 
RGB=insertShape(A_gray,'rectangle',[column row Tc Tr], 'LineWidth',3); %create a box around maximum match
imshow(RGB);

% RIGHT HERE I WOULD LIKE TO COMPARE THE OBJECTS WITH A NORMAL MODULE
% IN ORDER TO FIND DEFECT MODULES AND LOCATE THEIR POSITIONS

Answer 1

[Shunichi Kusano]

You select only one of highly correlated pixels, because you find the maximum correlation by "find(R==(max(max(R))))". In order to detect some of them, you need to select peaks. For that, the following link is useful.

How can I find the peaks of a surface

By the way, you can use max(R(:)) instead of max(max(R)). In addition, j = j + 1 in the for loop is not needed here. In the for loop, the value of j is automatically incremented according to what you indicated as 1:max(label(:)).

[Felipe Souza]

Thanks a lot!

I am going to give a look on the link you posted