How to count number of circles of a gray scale image?

I want to count number of circles of a gray scale image. I have attached the sample image. I just want to count the number of circle of that image and fill the gaps of that circle. What should i do??

Answer 1

[Image Analyst]

How many do you see? I see two circular black discs plus a bunch of other blobs that have irregular shapes, though some parts of their perimeters are circular.

See this link to split them apart, if you want to do that. http://www.mathworks.com/help/images/examples/marker-controlled-watershed-segmentation.html

Then you can count them with bwlabel();

[~, NumBlobs] = bwlabel(binaryImage);

If you want to, you can try imfindcircles. http://www.mathworks.com/matlabcentral/fileexchange/34365-circle-finder

http://blogs.mathworks.com/pick/2008/05/23/detecting-circles-in-an-image/

Another way is to use regionprops() to compute the area and perimeter of the blobs and compute the circularity and keep only blobs that are sufficiently close to being circular.

I can't do it all for you but if you have questions, attach whatever code you have.

[Ruhul Amin]

Basically i want to fill up those holes of circles and blobs and then i want to detach those blobs and circles using erosion. But when i am using imfill() to fill up those holes it fill up the intersected area of those circles. What should i do?

Here is my code:

if true
I = imread('sample.png');
% figure, imshow(I);title('Main Image');
BW = im2bw(I);
figure, imshow(BW);                                              
BW = imfill(BW,'holes');
figure, imshow(BW);
for n=1:5
  BW = bwmorph(BW,'erode');
end;
figure, imshow(BW);title('erode');  
end

[Image Analyst]

Split them apart first, then fill.

[Fika Umami]

this post is quite old but ca

n i ask here??

im looking fora solution tom problem but i haven't found it yet

i tried to calculate the circle from the canny result but the results did not match my manualy calculation, the results were too many

this my code

edge_final = handles.edge1;
cla('reset')
imshow(edge_final)

[B,L] = bwboundaries(edge_final,'noholes');
stats = regionprops(L,'All');
  
bw2 = zeros(size(edge_final,1),size(edge_final,2));
  
for k = 1:length(B)
    boundary = B{k};
    delta_sq = diff(boundary).^2;
    perimeter = stats(k).Perimeter;
    area = stats(k).Area;
    metric = 4*pi*area/perimeter^2;
    bbox = stats(k).BoundingBox;
    eccentricity = stats(k).Eccentricity;
    MajorAxisLength = stats(k).MajorAxisLength;
    MinorAxisLength = stats(k).MinorAxisLength;
    d = MajorAxisLength/MinorAxisLength;
    position = [boundary(1,2),boundary(1,1)];
      
    if metric > 0.9
        bw2(L==k) = 1;
        [handles.D handles.Num_lingkaran]=bwlabel(bw2);
        set(handles.text2,'string', handles.Num_lingkaran);
    end
    
end
guidata(hObject, handles);

can you help me to o it?

[Image Analyst]

I'd create your own post for this. Attach your own original input image.

Also explain exactly what "calculate the circle" means because that is very vague and there could be several possible answers for that depending on exactly what that means.