Getting distance of centroids from each other.

Is it possible to get the distances between each centroid? For example I have 5 centroids (centroid 1 - 5) in an image and I would like to get the distance between centroid 1 and centroid 2, then centroid 1 and centroid 3, then centroid 1 and centroid 4, then centroid 1 and centroid 5, then for centroid 2 and centroid 3, and so on. Then I will store these distances in an adjacency list or an adjacency matrix so that I could create a graph network.

Answer 1

[Walter Roberson]

use pdist()

[Sean Ivan Roxas]

I have this chunk of code already that prints a line between connected coins, but this only takes into account adjacent centroids not all. Can you please edit it such that it can also give me the distances between all the coins and store it in an adjacency matrix/list?

%% Blob Analysis

hBlobAnalysis = vision.BlobAnalysis('MinimumBlobArea', 10000,'MaximumBlobArea',150000);
[objArea,objCentroid,bboxOut] = step(hBlobAnalysis, IBWopen);

%% Print Centroid

CC = bwconncomp(IBWopen);
S = regionprops(CC, 'centroid');
centroids = cat(1, S.Centroid);

NodeCount = length(objCentroid)

figure
subplot(1,2,2)
imshow(pic)

hold on
plot(centroids(:, 1), centroids(:, 2), 'b*','LineWidth', 5)
hold off

%% Print Lines 

props = regionprops(IBWopen , 'Centroid', 'EquivDiameter');
meanDiam = mean([props.EquivDiameter]);
xy = vertcat(props.Centroid);
distances = pdist2(xy, xy);

hold on;
for k = 1 : size(distances, 1)
    thisRow = distances(k, :);
    closeIndexes = find(thisRow > 0 & thisRow < 1.12 * meanDiam);
   
    for k2 = 1 : length(closeIndexes)
        x1 = xy(k, 1);
        x2 = xy(closeIndexes(k2), 1);
        y1 = xy(k, 2);
        y2 = xy(closeIndexes(k2), 2);
        plot([x1, x2], [y1, y2], 'k-', 'LineWidth', 2)
    end
    
end
hold off;

[Walter Roberson]

distances = squareform(pdist2(xy)) ;

is equivalent to what you have; I am not sure which would be faster. Either way, distances would already contain the distance between each centroid and each other centroid.

Since the distance between each point and itself is 0, then the distances array already acts as a weighted adjacency matrix.

[Sean Ivan Roxas]

Alright, my question now is that would I be able to create a graph network using MATLAB out of this weighted adjacency matrix similar to python's NetworkX?

[Walter Roberson]

G = graph(distances)

[Sean Ivan Roxas]

Is it possible to use graphtraverse or dfsearch to calculate for the effective resistance of the graph?