How to generate random smooth 2D closed curves/shapes

Question

Roy Velich 2020-8-11

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/578254-how-to-generate-random-smooth-2d-closed-curves-shapes

回答： Moreno, M. 2022-4-15

Hi,

I would like to know how can I generate a collection of random 2D closed smooth curves. I thought about generating a random 3D surface with random peaks, and then intersecting the Z=0 plane with it, and extract the largest contour, for example. I just not sure how to do it practically... it is the correct way? I would love to get some guidance.

I am look for a result in this fashion:

Thanks

3 个评论
显示 1更早的评论隐藏 1更早的评论

Roy Velich 2021-2-4

Yes, I endded up downloading thousands of random images from the web, applying Guassian blur on each of them (with a collection of kernels) and extracting closed level curves from the blurred images. I can give you my code if you want.

Roy Velich 2021-2-4

https://github.com/xenomarz/smooth-curves-generator

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

hosein Javan 2020-8-12

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/578254-how-to-generate-random-smooth-2d-closed-curves-shapes#answer_478482

If you are familiar with bezier curve, you can do it easily by setting random values for points coordinates.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 2

Moreno, M. 2022-4-15

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/578254-how-to-generate-random-smooth-2d-closed-curves-shapes#answer_943910

在 MATLAB Online 中打开

The function bspl is creation of my own and is available in these 2 links:

https://uk.mathworks.com/matlabcentral/fileexchange/109179-high-precision-bezier-curves?s_tid=srchtitle

https://uk.mathworks.com/matlabcentral/fileexchange/109184-high-precision-bezier-curve-fitting-and-modelling?s_tid=srchtitle

Given a closed Bézier polygon at which the initial and final points are coincident, and the adjacent points to the ends of the curve are colinear, the generated Bézier curve is closed and smooth. If you want your curve not to self-intersect, then your initial set of data points should be the bounding points of a point cloud. If the adjacent points to the end of the curve are equidistant, then your curve has C3 continuity:

tiledlayout(2, 2)
for i = 1 : 4
    X = rand(10, 2);
    X = X(boundary(X), :);
    X(length(X) - 1, :) = 2 * X(1, :) - X(2, :);
    X = bspl(X, 1e3);
    nexttile
    plot(X(:, 1), X(:, 2))
    axis equal
end