How to extract connected chains in canny edge image? Specifically asking, how would I extract connected chains when there are so many edges are available, so that I could come up with top 10 results (top 10 long edges).

Question

Jay 2016-3-20

评论： Jay 2016-3-22

Answer 1

Matthew Eicholtz 2016-3-21

Would regionprops do the trick?

----

Read the image:

I = imread('cameraman.tif');

Compute edges and create label matrix:

J = edge(I,'canny');
L = bwlabel(J);

Extract properties for connected components:

s = regionprops(J,'PixelIdxList');

Sort based on number of pixels in each connected component:

d = cellfun('length',{s(:).PixelIdxList}); %total number of pixels in each region
[~,order] = sort(d,'descend');

Show top-10 connected components:

K = ismember(L,order(1:10)); %only show the top 10 edges

Jay 2016-3-22

thanks. Do you have any idea how would I fill the gaps in contours in canny image before extracting connected chains?

It may give more promising results, I guess.