Getting Started with Triangle.NET in C# and .NET When working with computer graphics, geographic information systems (GIS), or finite element analysis (FEA), you often need to break down complex 2D shapes into triangles. This process is called triangulation. Triangle.NET is a powerful, open-source C# port of Jonathan Shewchuk’s robust Triangle library, specifically designed to handle Delaunay triangulations, Voronoi diagrams, and high-quality mesh generation.
Here is how to get started with Triangle.NET in modern .NET applications. Why Choose Triangle.NET?
Speed and Reliability: It uses highly optimized geometric predicates to prevent numerical errors and crashes caused by precision issues.
Constrained Triangulation: You can force the mesh to respect specific boundaries, lines, or structural borders.
Refinement Capabilities: It allows you to specify maximum triangle areas or minimum angles to create high-quality, uniform meshes. 1. Installation
To begin, add the Triangle.NET package to your C# project. You can install it via the NuGet Package Manager Console: Install-Package TriangleNet Use code with caution. Or using the .NET CLI: dotnet add package TriangleNet Use code with caution. 2. Basic Delaunay Triangulation
A standard Delaunay triangulation connects a set of points such that no point falls inside the circumcircle of any triangle in the network. This maximizes the minimum angles of the triangles, avoiding thin, stretched shapes. Here is how to triangulate a simple cloud of random points:
using TriangleNet.Geometry; using TriangleNet.Topology; class Program { static void Main() { // 1. Create a geometry container var polygon = new Polygon(); // 2. Add random cloud points polygon.Add(new Vertex(0.0, 0.0)); polygon.Add(new Vertex(5.0, 0.0)); polygon.Add(new Vertex(2.5, 5.0)); polygon.Add(new Vertex(2.5, 2.0)); // Interior point // 3. Triangulate the points var mesh = polygon.Triangulate(); // 4. Iterate and display the generated triangles Console.WriteLine(\("Generated {mesh.Triangles.Count} triangles:"); foreach (var triangle in mesh.Triangles) { var p0 = triangle.GetVertex(0); var p1 = triangle.GetVertex(1); var p2 = triangle.GetVertex(2); Console.WriteLine(\)“Triangle: ({p0.X}, {p0.Y}) -> ({p1.X}, {p1.Y}) -> ({p2.X}, {p2.Y})”); } } } Use code with caution. 3. Constrained Delaunay Triangulation (CDT)
In real-world applications, you rarely triangulate random points. Usually, you have a specific boundary—like a game map or a mechanical part outline—that the triangles must not cross. This requires Constrained Delaunay Triangulation.
By adding segments to your polygon, you tell Triangle.NET to treat those lines as solid walls:
var polygon = new Polygon(); // Define boundary vertices polygon.Add(new Vertex(0, 0), 0); polygon.Add(new Vertex(10, 0), 1); polygon.Add(new Vertex(10, 10), 2); polygon.Add(new Vertex(0, 10), 3); // Define segments to lock the outer boundary layout polygon.Add(new Segment(polygon.Vertices[0], polygon.Vertices[1])); polygon.Add(new Segment(polygon.Vertices[1], polygon.Vertices[2])); polygon.Add(new Segment(polygon.Vertices[2], polygon.Vertices[3])); polygon.Add(new Segment(polygon.Vertices[3], polygon.Vertices[0])); // Triangulate respecting the segments var mesh = polygon.Triangulate(); Use code with caution. 4. Mesh Quality and Refinement
If you are running physics simulations, large or highly skewed triangles can cause inaccurate results. You can pass configuration options to Triangle.NET to refine the mesh automatically, breaking large triangles into smaller, more uniform ones.
using TriangleNet.Meshing; // Define quality options var options = new QualityOptions { MaximumArea = 2.0, // No triangle can be larger than 2.0 units MinimumAngle = 25.0 // Try to keep all angles above 25 degrees }; // Generate a high-quality refined mesh var mesh = polygon.Triangulate(options); Use code with caution. 5. Adding Holes to Your Mesh
If you need to model a shape with a cutout—such as a donut shape or a room with an obstacle—you can define a closed loop of segments inside your polygon and add a “Hole” point. Triangle.NET will automatically hollow out the area surrounding that point.
// 1. Define outer boundary (Vertices 0 to 3) and segments… // 2. Define inner hole vertices polygon.Add(new Vertex(3, 3), 4); polygon.Add(new Vertex(7, 3), 5); polygon.Add(new Vertex(7, 7), 6); polygon.Add(new Vertex(3, 7), 7); // 3. Connect hole segments polygon.Add(new Segment(polygon.Vertices[4], polygon.Vertices[5])); polygon.Add(new Segment(polygon.Vertices[5], polygon.Vertices[6])); polygon.Add(new Segment(polygon.Vertices[6], polygon.Vertices[7])); polygon.Add(new Segment(polygon.Vertices[7], polygon.Vertices[4])); // 4. Place a Hole marker inside the inner segment boundary polygon.Add(new Hole(5.0, 5.0)); // 5. Triangulate (the center square will remain empty) var mesh = polygon.Triangulate(); Use code with caution. Conclusion
Triangle.NET takes the headache out of 2D computational geometry in C#. Whether you need a simple point-cloud mesh or a highly refined structural grid with complex holes and borders, its straightforward API delivers fast and mathematically robust results. Experiment with different QualityOptions to find the perfect balance between triangle count and shape quality for your specific application.
If you would like to expand this implementation, let me know if you want to focus on: Exporting data to formats like OBJ or DXF Generating Voronoi diagrams from your mesh Rendering the mesh visually using GDI+, WPF, or MonoGame
Leave a Reply