Plot polygons as image to process

Hi guys,

I have data of a curve which I want plot it as an image. And later use function regionprops to get its orientation and centroid.

Is there any way I can do that, or directly get the orientation of the curve data.

The image is the polygon plot of the curve, while the data has not duplicate point, the function return:

"Warning: Polyshape has duplicate vertices, intersections, or other inconsistencies that may produce inaccurate or

unexpected results. Input data has been modified to create a well-defined polyshape."

And delete half of my data to draw that image, any suggestion I could do to prevent that?

Thanks!

Answer 1

[Simon Chan]

You may use function centroid to determine the centroid of a polyshape.

When you convert the polyshape into an image, the result of the centroid would be different due to the resolution of a pixel on an image.

Regarding to the warning message, you may refer to this: https://www.mathworks.com/matlabcentral/answers/606836-why-does-the-polyshape-function-deletes-the-vertices-in-the-input-when-there-are-no-duplicated-poi

Consider the following example with a very simple polyshape:

pgon = polyshape([20 20 17 15],[7 3 2 9]);
[x,y] = centroid(pgon)

x = 17.7984

y = 5.4419

Now, convert it to an image:

A = zeros(25,25);
for k = 1:length(pgon.Vertices)
    A(pgon.Vertices(k,2),pgon.Vertices(k,1))=1;
end
J = bwconvhull(A);
regionprops(J,'Centroid','Orientation')

ans = struct with fields:
       Centroid: [17.6875 5.5000]
    Orientation: 63.9994

If you want to increase the resolution, for example, 100 times. This would give you a closer value.

MagFactor = 100;
A = zeros(25*MagFactor,25*MagFactor);
for k = 1:length(pgon.Vertices)
    A(pgon.Vertices(k,2)*MagFactor,pgon.Vertices(k,1)*MagFactor)=1;
end
J = bwconvhull(A);
stat = regionprops(J,'Centroid','Orientation')

stat = struct with fields:
       Centroid: [1.7798e+03 544.2466]
    Orientation: 65.4331

cent = stat.Centroid/MagFactor

cent = 1×2

17.7977 5.4425

figure
ax1 = subplot(1,2,1);
plot(pgon);
axis(ax1,'image');
ax1.XLim = [1 25];
ax1.YLim = [1 25];
ax2 = subplot(1,2,2);
imshow(J);
ax2.YDir='normal';

[Nhat Nguyen]

Fantastic,

Thank you mate!

[Eric]

Very helpful, many thanks

[Eric]

insertShape seems to be the standard Matlab solution...

[Image Analyst]

@Eric, insertShape burns a shape into an image array. It requires the Computer Vision Toolbox.

poly2mask creates a binary image and requires the Image Processing Toolbox. It does not burn pixels/shapes into another image though you could use the binary image to burn it into a gray scale or RGB image.

Functions fill and patch work on plots (not images, though you can put hold on and use them over images, in which case they'd be in the overlay, not burned into the image itself). They are included in standard, base MATLAB.

Answer 2

[Image Analyst]

If you have the x and y of the outline, you can use poly2mask to turn it into an image and then use regionprops to get its orientation and centroid.

mask = poly2mask(x, y, rows, columns);
props = regionprops(mask, 'Orientation', 'Centroid');