Hello Surendra,
But the general idea would be to:
- Crop ROI
- Segment your image
- Obtain the boundaries and calculate thickness
clear all
close all
I=imread('Original_image (2).png');
%% Crop region
Icrop=I(1:720,400:600);
%% Segment
BW=~imbinarize(Icrop);
se = strel('disk',5);
BWc=imclose(BW,se);
skelImage = bwskel(BWc, 'MinBranchLength', 10); % First guess. Strangely bwskel() doesn't seem to have any short spurs, no matter what MinBranchLength is.
imshow(imoverlay(BWc, skelImage, 'r'));
title('Binary image with skel')
%% Obtain distances
edtImage = bwdist(~BWc); %Distance of to nearest 0 value
radii = edtImage(skelImage); % Extract the distances only along the skeleton
% These are the half widths. Multiply by 2 to get the full widths.
averageWidth = 2 * mean(radii)
%% Find the boundaries
E=edge(BWc); % Find the edges
CC = bwconncomp(E); % Isolate the different components
[rowOut1,colOut1]=ind2sub(size(E),CC.PixelIdxList{1,1});
Pout1=[rowOut1,colOut1]; % Pixels forming Outer boundary 1
[rowIn,colIn]=ind2sub(size(E),CC.PixelIdxList{1,2});
Pin=[rowIn,colIn]; % Pixels forming Inner boundary 1
[rowOut2,colOut2]=ind2sub(size(E),CC.PixelIdxList{1,3});
Pout2=[rowOut2,colOut2]; % Pixels forming Outer boundary 2
figure(2)
imshow(BWc)
hold on
plot(Pin(:,2),Pin(:,1),'.','MarkerSize',2)
plot(Pout1(:,2),Pout1(:,1),'.','MarkerSize',2)
plot(Pout2(:,2),Pout2(:,1),'.','MarkerSize',2)
title('Boundaries')
Hope this helps
