How to find the center co-ordinates of different regions

Hi there,

Can anyone assist me how to find the center co-ordinates automatically and save the co-ordinate (x and y position) values in an array automatically? The shapes are different for the regions but each region has same pixel value.

Please have a look at my image and guide me how to find the center co-ordinates of each region and save them in an array automatically without using manual operation.

Any help will be highly appreciated.

Answer 1

[Image Analyst]

I thought you'd seen it before, but I guess not. See my image segmentation tutorial in my File Exchange http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862

Basically it's something like

grayImage = rgb2gray(rgbImage);
[labeledImage, numberOfBlobs] = bwlabel(grayImage ~= 0);
measurements = regionprops(labeledImage, 'Centroid');
allCentroids = [measurements.Centroid];
xCentroids = allCentroids(1:2:end);
yCentroids = allCentroids(2:2:end);

[Tania]

Thank you very much however; I am only getting 4 co-ordinates from the image. No clue why is that like this:

xCentroids =

   96.0000  228.5272  362.5419  415.0248

yCentroids =

   61.0000  120.2121  206.5058   70.5421

I have tried the other way using

s = regionprops(BW,'centroid');
centroids = cat(1,s.Centroid)

centroids =

   96.0000   61.0000
  228.5272  120.2121
  362.5419  206.5058
  415.0248   70.5421

Could you please tell me why it is giving me only 4 co-ordinates instead of 6?

Thank you in advance.

[Guillaume]

I only count 5 shapes in your image.

In any case, check that your binary image has indeed 5 (or 6) regions in it. Most likely, your segmentation has gone wrong.

[Tania]

Now I am getting 5 co-ordinates. Should not I get 6 (should not I consider black region as well) instead of 5?

centroids =

   98.5119   81.5119
  167.4810  263.8270
  230.5137  121.0732
  364.5173  208.5051
  417.0518   72.5668

Not sure what I have done wrong here.

Please advice me. Thanks

[Image Analyst]

You didn't follow my recommendations, that's what. I said to do this:

grayImage = rgb2gray(rgbImage);

I did not say to use im2double() or im2bw(). Also, do not use image as the name of your filename string. image is the name of a built-in function. Call it fullFileName or something like that.

[Tania]

I have figured it out to get 5 regions so far. I can not figure out how to get the black one.

If you can advice me then it would be great.

I appreciate your time and help.

Thank you very much.

[Image Analyst]

If you want to treat the black background as its own object, then you'll have to do that separately (well, that's the simplest and most straightforward way).

grayImage = rgb2gray(rgbImage);
[labeledImage, numberOfBlobs] = bwlabel(grayImage == 0);
measurements = regionprops(labeledImage, 'Centroid');
allCentroids = [measurements.Centroid];
xCentroids = allCentroids(1:2:end);
yCentroids = allCentroids(2:2:end);

Note the == 0 instead of ~= 0 so that we pick up only pure black pixels.

[Tania]

Thank you very much. That solved my problem.

Great help