Swiftpack.co -  Package - CooperCorona/Voronoi
Swiftpack.co is a collection of thousands of indexed Swift packages. Search packages.
Swift framework for calculating voronoi diagrams using Fortune's Algorithm.
.package(url: "https://github.com/CooperCorona/Voronoi.git", from: "2.0.0")


Swift framework for calculating voronoi diagrams using Fortune's Algorithm.

Use the VoronoiDiagram class to calculate the cells of a voronoi diagram. Voronoi diagrams are the edges that bisect the lines between given points. Each point within a cell is closer to that cell's voronoi point than any other voronoi point. Points lying on the edges of two cells are equidistant to the two voronoi points.

Fortune's Algorithm

Fortune's Algorithm is a way of solving voronoi diagrams in O(n log(n)) time. It only needs to process an event at each voronoi point and at possible circles formed by three voronoi points (which occurs in O(n) time), and also needs to search a binary tree at each site event (which occurs in O(log(n)) time).

Fortune's algorithm uses a beach line (a piecewise curve formed by the minimum value of parabolas at a given x-coordinate) and a sweep line (a horizontal line corresponding to the directrix's of the parabolas). Each voronoi point corresponds to the focus of a parabola and the sweep line corresponds to the directrix.

A parabola can be defined as a focus (a point) and a directrix (a line). The distance between any point on the parabola and its focus is equal to the distance between that same point and the directrix (which, given that the directrix is a horizontal line, is just the difference in y-coordinates). For example, suppose our parabola has focus (x, y) = (0, 1) and directrix y = -1. (1, 0.25) lies on the parabola. The distance between the focus and that point is sqrt((0 - 1)^2 + (1 - 0.25)^2) = sqrt(1 + 0.5625) = sqrt(1.5625) = 1.25 and the distance between the point and the directrix is 0.25 - (-1) = 1.25, which is the same.

Since the distance between a point and the focus and that same point and the directrix is the same, the intersection of two parabolas with the same directrix (the sweep line) forms the bisector between those two points. By adding new parabolas to the beach line (and splitting the previous parabolas) at every new voronoi point, Fortune's algorithm traces the edges of the voronoi diagram.

The final consideration is circle events. When three focii all lie on the same circle, that means all three parabolas will intersect when the sweep line reaches the top of the circle, which means the middle parabola is "squeezed" out of the beach line. Thus, when we add or remove parabolas from the beach line, we check if they form circle events. A circle event is also where two edges meet and a new one begins.

Once all events are processed (and incomplete edges are extended to the boundaries), the voronoi diagram is complete.

Voronoi Cells

In some cases, you want to know more than the edges of a voronoi diagram. After the voronoi diagram is complete, you can calculate the vertices associated with each individual voronoi point. This is useful, for example, if you're trying to render a filled-in voronoi diagram and you need to know the vertices of a cell to form triangles. The VoronoiResult class exposes an array of VoronoiCell objects, which themselves allow you to calculate the vertices of a given cell (because the vertices are not calculated until you explicitly invoke the method, there is no performance penalty if you do not need to access them).


  • Add images of completed voronoi diagrams.


Stars: 19
Last commit: 2 weeks ago

Related Packages

Swiftpack is being maintained by Petr Pavlik | @ptrpavlik | @swiftpackco