How can imfill a objects when they are in the edges of main image?

39 views (last 30 days)
hi,
I have Two rectangles 200X300 and 300X100 in the edges as shown below. I need to imfill in side the objects edge shape only. I have try use the regular imfill but doesn't work with me. Could you please help me??p.jpg
  4 Comments

Sign in to comment.

Accepted Answer

Image Analyst
Image Analyst on 26 Nov 2019
Edited: Image Analyst on 24 Nov 2020
Close the contours then fill. Assuming the lines reach the edge lines:
% Make top line white:
col1 = find(binaryImage(1, :), 1, 'first')
col2 = find(binaryImage(1, :), 1, 'last')
binaryImage(1, col1:col2) = true;
% Make bottom line white:
col1 = find(binaryImage(end, :), 1, 'first')
col2 = find(binaryImage(end, :), 1, 'last')
binaryImage(end, col1:col2) = true;
% Make right line white:
row1 = find(binaryImage(:, end), 1, 'first')
row2 = find(binaryImage(:, end), 1, 'last')
binaryImage(row1:row2, end) = true;
binaryImage = imfill(binaryImage, 'holes');
If the lines don't reach the edges, then you'd need to use bwmorph() to find endpoints, then use bwlabel() to identify separate/distinct blobs, then construct a line between the two end points of each blob using polyfit(). Then write those lines into the image, then finally call imfill(). This is a more complicated algorithm but certainly doable in a few lines of code.
Post your original image (not a screenshot) if you still have trouble.
  2 Comments
Image Analyst
Image Analyst on 24 Nov 2020
Santiago: Sorry, I must have gotten confused. There never was an imfind(). It should be find(). I must just be used to putting im in front because so many of the Image Processing Toolbox functions have im in front of them. Thanks for the catch and I'll make the correction.

Sign in to comment.

More Answers (2)

KALYAN ACHARJYA
KALYAN ACHARJYA on 26 Nov 2019
Edited: KALYAN ACHARJYA on 26 Nov 2019
Its because of white area touches the border. Apply imclearborder on binary image, then only apply imfill

Turlough Hughes
Turlough Hughes on 26 Nov 2019
You could put a border on the image with value 1, then run the filter and remove the border to get original dimensions:
I=padarray(testiamge,[1 1],1,'both');
test=edge(I,'canny');
outputs = imfill(test,'holes');
outputs([1 end],:)=[]; outputs(:,[1 end])=[];
imshow(outputs)
Also, this might also be an option depending on your requirements:
outputs=~testimage;

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!