Multiple line fitting in binary image

Dear colleagues,

I would like to kindly ask you on the issue with multiple line fitting in MATLAB. I have a binary image with multiple skeletons, and I need to perform fitting of each line in this image by continous curve. I am able to do that for one line, see the code. Nevertheless, I am unable to do that for the all the lines at once automaticaly and plot the fitting results.

I use the image labeling: bwlabel() to recognize individual lines and for the fitting, I use the function smooth().

I would like to kindly ask you for the help with this task.

Thank you very much in advance.

I use the following code:

i=imread('test.png');

image=i(:,:,2);

BW = logical(image);
BW1=bwskel(BW);
[skel_label N]=bwlabel(BW1);

figure()
imshow(BW1)

[r s]=size(skel_label);

for ii=1:r
    for jj=1:s
        if skel_label(ii,jj)==2
            skel_label_mod(ii,jj)=1;
        else
             skel_label_mod(ii,jj)=0;
        end
    end
end

[coor1 coor2]=find(skel_label_mod);

yy1 = smooth(coor2,coor1,0.5,'loess');

figure()
imshow(i)
hold on
plot(coor2,yy1,'x')

Answer 1

[Harikrishnan Balachandran Nair]

Hi,

I understand that you are trying to use the 'smooth' function on each of the connected components/curved lines that you got using the 'bwskel' function.

Using the 'bwlabel' fuction, each of the connected component in the corresponding image can be given a label number. This label can be used with the 'find' function to smooth the line with the corresponding label number. You can refer to the following code.

[skel_label N]=bwlabel(BW1); 
figure()
imshow(i)
hold on;
for i=1:N
    [coor1 coor2]=find(skel_label==i); % Finding all the co-ordinates for the corresponding component
    yy1 = smooth(coor2,coor1,0.5,'loess');
    plot(coor2,yy1,'x');
end
hold off;