MATLAB: Find pixel coordinates value of a centroid

image segmentation

[EDIT: Fri May 13 20:58:34 UTC 2011 – Reformat – MKF]

I want to find the x n y coordinates and pixel value of the centroid. Here is the code I have:

I = imread('mobil.png');
%I2 = imtophat(I,strel('disk',15));
level = graythresh(I);
bw = im2bw(I,level);
I2 = 1 - bw;
I2 = bwareaopen(I2, 200);
I3 = imfill(I2, 'holes');
s  = regionprops(I3, 'centroid');
centroids = cat(1, s.Centroid);
imshow(I3)
hold on
plot(centroids(:,1), centroids(:,2), 'b*')
hold off
impixelinfo

The car is :

Image

Best Answer

centroids = round(centroids);
cent_vals = Inotbw(sub2ind(size(Inotbw),centroids(:,2),centroids(:,1)))

Related Solutions

MATLAB: How can i separate all the connected regions in a binary image and display them individually.

Use bwlabel() and ismember

[labeledImage, numRegions] = bwlabel(binaryImage);
props = regionprops(labeledImage, 'Area')
% Compute all the areas.
allAreas = [props.Area]
% Display each blob in its own figure.
for k = 1 : numRegions
    thisRegion = ismember(labeledImage, k);
    figure;
    imshow(thisRegion);
    drawnow;
end

MATLAB: How i know the value of a pixel in a centroid of image

Easy way: round the coordinates returned from regionprops and use those as indices to get the centroid. Harder way is to interpolate that point based on its 4 nearest neighbors.
You absolutely can do it all in one m-file. It really depends on the application whether you want to or not. If it's a simple one line threshold (eg. bw = x>50) then definitely do it in one file. If it's something more complicated (eg. distance to the nearest location on the convex hull of an image of connected tauri) then you probably want a second m-file designed specifically for that.

Best Answer

Related Solutions

MATLAB: How can i separate all the connected regions in a binary image and display them individually.

MATLAB: How i know the value of a pixel in a centroid of image

Related Question