Plotting dots on the edges of an object uniformly

Hello, everyone. I wonder is it possible to distribute a specific number of dots on the edges of the acquired image. Think of a circle for instance, is it possible to distribute,like, 8 dots uniformly on the circle ? Note: As in the image i have uploaded. Red circles are the dots etc.(sorry for paint level image :3)

Answer 1

[Walter Roberson]

See http://www.mathworks.com/matlabcentral/answers/317921-distances-along-a-plotted-line#answer_248275

You would probably want to set marker_dist to dist_from_start(end)/N where N is the number of dots you want around the edge.

Note that that code does not try to put the markers at corners: it is code for placing the markers at equal distances along the path, where-ever that comes out to be.

[Kerem Asaf Tecirlioglu]

hello thanks for the answer. But there is an issue it seems quite basic but i couldn't solve it . I get an error message,

Index exceeds matrix dimensions. Error in

marker_x = x(marker_base_pos) .* (1-weight_second) + x(marker_base_pos+1) .* weight_second;

    x = [0 5 10];
     y = [2 4 1];
     dist_from_start = cumsum( [0, sqrt((x(2:end)-x(1:end-1)).^2 + (y(2:end)-y(1:end-1)).^2)] );
     a=dist_from_start(end)/5;
     marker_dist=a;
     marker_locs = marker_dist : marker_dist : dist_from_start(end);   %replace with specific distances if desired
     marker_indices = interp1( dist_from_start, 1 : length(dist_from_start), marker_locs);
     marker_base_pos = floor(marker_indices);
     weight_second = marker_indices - marker_base_pos;
     marker_x = x(marker_base_pos) .* (1-weight_second) + x(marker_base_pos+1) .* weight_second;
     marker_y = y(marker_base_pos) .* (1-weight_second) + y(marker_base_pos+1) .* weight_second;
     plot(x, y);
     hold on;
     plot(marker_x, marker_y, 'r+');
     hold off

[Walter Roberson]

Hmmm, I remember fixing that at the time, but I do not remember what the fix I used then was. One fix would be

     x = [0 5 10];
     y = [2 4 1];

     Number_of_divisions = 5;

     dist_from_start = cumsum( [0, sqrt((x(2:end)-x(1:end-1)).^2 + (y(2:end)-y(1:end-1)).^2)] );

     a = dist_from_start(end)/Number_of_divisions;
     marker_dist = a;

     marker_locs = marker_dist : marker_dist : dist_from_start(end);   %replace with specific distances if desired

     marker_indices = interp1( dist_from_start, 1 : length(dist_from_start), marker_locs);
     marker_base_pos = floor(marker_indices);
     weight_second = marker_indices - marker_base_pos;

     mask = marker_base_pos < length(dist_from_start);
     final_x = x(end); final_y = y(end);
     final_marker = any(~mask) & (x(1) ~= final_x | y(1) ~= final_y);

     marker_x = [x(1), x(marker_base_pos(mask)) .* (1-weight_second(mask)) + x(marker_base_pos(mask)+1) .* weight_second(mask), final_x(final_marker)];
     marker_y = [y(1), y(marker_base_pos(mask)) .* (1-weight_second(mask)) + y(marker_base_pos(mask)+1) .* weight_second(mask), final_y(final_marker)];

plot(x, y, marker_x, marker_y, 'r+');

This code also checks to see if the curve is (exactly) closed and in that case avoids plotting the marker twice (because there is already a marker there from the plot from the beginning.)

Answer 2

[Image Analyst]

See interparc http://www.mathworks.com/matlabcentral/fileexchange/34874-interparc by John D'Errico. It works for arbitrarily shaped curves in 2D as well as circles.

[Walter Roberson]

Note: a difference between my code and John's is that my code treats the points as a series of line segments that the markers must sit on, whereas John's code treats the points a subsampling of a curve and interpolates the curve, a more difficult task.

My code is suitable for the case where the points define the path, such as the situation of a vehicle traveling at constant velocity along a map.