How can I calculate area of an object in the image with .png extension?

Hi there,

I want to obtain area of an object in the png image. Could anyone please tell me how to do it? One more thing, is there any way to calculate length a between 2 points of the horizontal line and the curve? (as picture)

I'm not sure there have been answers for my question, although I've been trying to find them on the internet. But they don't work in my situation.

Any help would be greatly appreciated. Excuse for my English.

Khanh.

[Kiran Das]

What will the area of the object in this case?

Answer 1

[Rushikesh Tade]

I have converted the image into problem image by removing the text which you used to describe the problem and filling the background with black color.

Following is the code to do it.

input_im=imread('area2.png');
input_im=rgb2gray(input_im);

imshow(input_im)
title('Original Object')

After that I have tried separating the object out:

sum_of_x_axis=sum(input_im,1);
sum_of_y_axis=sum(input_im,2);

location_of_object_on_x=find(sum_of_x_axis>0);

location_of_object_on_y=find(sum_of_y_axis>0);    

seperated_object=input_im(location_of_object_on_y(1):...
    location_of_object_on_y(end),location_of_object_on_x(1):...
    location_of_object_on_x(end));

figure;
imshow(seperated_object);
title('Seperated Object')

Now the area of the object can be calculated as : (seperated_object_image_area) - (area_which_does_not_belong_to_object_in_the_seperated_object_image)

and Height of the line ("a" displayed in your image) can be found as:maximum vertical sum along the x axis.

Following is the code for that:

detected_outer_surface=im2bw(seperated_object);
sum_of_outer_space=0;
for i=1:size(seperated_object,1)
   if seperated_object(i,1)>0
       break
   end
    for j=1:size(seperated_object,2)
         if seperated_object(i,j)>0
        break
         end
         detected_outer_surface(i,j)=1;
         sum_of_outer_space=sum_of_outer_space+1;
    end
end

for i=1:size(seperated_object,1)
   if seperated_object(i,size(seperated_object,2))>0
       break
   end
    for j=size(seperated_object,2):-1:1
         if seperated_object(i,j)>0
        break
         end
         detected_outer_surface(i,j)=1;
         sum_of_outer_space=sum_of_outer_space+1;
    end
end

figure;
imshow(detected_outer_surface)
title('Detected outer Surface which does not belong to object')

Area_of_object=size(seperated_object,1)*size(seperated_object,2);

Area_of_object=Area_of_object-sum_of_outer_space

Height_of_the_line=max(sum_of_x_axis)

Answers which I have got are:

   %Area_of_object =103353
   %Height_of_the_line =24240

[Khanh]

Thank you so much.

This is what I were looking for answers for my question.

This image was exported to png by AutoCAD. But the are of the object in the image calculated by your program is too different to its when I calculated by using "area" command in AutoCAD.

Could you please tell me why?

[Khanh]

I also attached the zip file including the dwg file. Please take a look. Thanks

[Khanh]

In addtion, the area of this object calculated by AutoCAD is 8669279.97 mm^2

[sadique sheikh]

Thank you Rushikesh. But this algorithms or program doesn't separate object from background in my attached file. What about my attached file? Please reply.

[sadique sheikh]

Thank you Rushikesh. But this algorithms or program doesn't separate object from background in my attached file. What about my attached file? Please reply.

[Image Analyst]

Twice now you have not attached a file. Are you using the paper clip icon? Maybe you should start your own, new discussion.

[Kiran Das]

Respected sir/madam,

can you show the algorithm of the above code?It will be very helpful to me.

Thank you.

[Image Analyst]

@Kiran Das. The code is there. That is the algorithm. If you want a tutorial on the basics of image analysis, see my Image Processing Tutorial on my File Exchange:

Image Analyst's File Exchange

[Kiran Das]

@Image Analyst Thank you for amswering my comment. But it is my humble request you to please explain the code for me. I can not get it as I am a begginer. This code is helpful to me it will be more fruitful to me as if someone expain the code.

Thank You.

[Image Analyst]

The code in this answer doesn't look like very good code (it's not my code). There are much better ways to do what he did. I suggest you don't try to understand it and just look at the Image Segmentation Tutorial in my File Exchange. It has lots of comments to walk/talk you through the basic image segmentation process.

Image Analyst's File Exchange

Answer 2

[Image Analyst]

Three steps.

1. Convert your color image to grayscale
2. Threshold the grayscale image
3. Either sum the white pixels, or call bwarea()

% Calculate the area using bwarea().
area1 = bwarea(binaryImage)
% Calculate the area in pixels using sum()
area2 = sum(binaryImage(:))

The attached code (full blown demo test.m attached in blue below the image) calculates the area both ways and produces this image.

I get 103353, or 103374.4 depending on the calculation method used.

[Image Analyst]

I think the bwarea method is similar to what you'd get if you used polyarea(). It gives you fractional pixels if you go along a diagonal because it doesn't count pixels but goes from "pixel center" to "pixel center".

[Khanh]

Thank you so much.

This is what I were looking for answers for my question.

This image was exported to png by AutoCAD. But the are of the object in the image calculated by your program is too different to its when I calculated by using "area" command in AutoCAD.

Could you please tell me why? And how to obtain the area of the object in the image same to its by using AutoCAD "area" command.

[Khanh]

I also attached the zip file including the dwg file. Please take a look. Thanks

[Khanh]

In addtion, the area of this object calculated by AutoCAD is 8669279.97 mm^2

[Image Analyst]

My answer is in pixels. You need to spatially calibrate to get a conversion factor that you can multiply the pixels by to get mm or mm squared. See attached demo.

[Khanh]

Your helps are so amazing. Thank you. I will study your comments and programs for my problem.

[Image Analyst]

You're welcome. You can always vote for good answers even though they were not the one answer you officially "Accepted". By the way, you can get the height of the line in a single line of code rather than those two dozen lines of 4 nested for loops in the answer you accepted.

height = find(binaryImage(:,col), 1, 'last')-find(binaryImage(:,col), 1, 'first');

where col is the column of the image where you're interested in knowing the height for.