Mechaide - Matlab Mesh 2D - DIMEC - UChile

www.mechaide.com

MATLAB tool for generating meshes from DXF files in 2D. Supports Delaunay triangulation and optimized polygonal meshes D-Polylla with Lloyd relaxation for VEM.

Requirements

• MATLAB (R2018b or later recommended).
• 2D DXF files with regions only, support regions with holes.

Installation

1. Clone the repository:

   git clone https://github.com/mechaide/matlab_mesh_2D.git 

Usage

Run the following MATLAB script to generate a mesh from a DXF file and automatically export it as mesh.json.

clear; clc;

% Parameters
dxf_file = 'myfile.dxf'; % Path to DXF file
d = 0.5; % Vertex separation
mesh_type = 'dpolylla'; % Mesh type: 'triangle' or 'dpolylla'
decimales = 2; % Decimal places for coordinate rounding
iterations = 3; % Lloyd relaxation iterations
plot_contour = true; % Plot contour
plot_base_mesh = true; % Plot base mesh
plot_frontier = true; % Plot frontier edges (dpolylla)
plot_contour_vertices = true; % Plot contour vertices
plot_mesh = true; % Plot final mesh

% Generate mesh
mesh_data = mechaide_MESH.generateMeshFromDXF(dxf_file, d, mesh_type, ...
    decimales, iterations, plot_contour, plot_base_mesh, plot_frontier, plot_mesh);

Parameters

• dxf_file: String. Path to DXF file (e.g., 'examples/test3.dxf'). Must contain valid lines/polylines.
• d: Double (positive). Max vertex spacing and min point distance for Poisson Disk Sampling. E.g., 0.5. Smaller values increase detail but computational cost.
• mesh_type: String ('triangle' or 'dpolylla'). 'triangle' for fast Delaunay triangulation; 'dpolylla' for optimized polygonal mesh (slower).
• decimales: Integer (non-negative). Decimal places for coordinate rounding. E.g., 2. Higher values increase precision but may reduce performance.
• iterations: Integer (non-negative). Number of Lloyd relaxation iterations. E.g., 3. More iterations improve mesh uniformity but increase computation time.
• plot_contour: Boolean. true to plot the DXF contour. Useful for geometry verification.
• plot_base_mesh: Boolean. true to plot the base mesh before relaxation.
• plot_frontier: Boolean. true to plot frontier edges (applies only to 'dpolylla').
• plot_contour_vertices: Boolean. true to plot contour vertices for debugging.
• plot_mesh: Boolean. true to plot the final mesh.

Sample Outputs

Below are example visualizations:

Output

• mesh_data: Struct with:
  • coords: Vertex coordinates (Nx2 matrix).
  • connect: Mesh connectivity (Mx3 matrix for triangles or cell array for polygons).
  • contorno: Contour points (matrix).
  • hoyos: Holes (cell array of point matrices).
  • Additional fields based on mesh_type.

Acknowledgments

Inspiration for this project comes from VEMLab by aaortizb, which provided insights into virtual element methods (VEM). The inspiration for this project comes from VEMLab by aaortizb, a project that contains the code to solve 2D static solids with FEM and VEM. The Direct Polylla (D-Polylla) algorithm mesh type in this project is based on the Polylla-Mesh-DCEL algorithm by ssalinasfe. Their work on polygonal meshing using terminal-edge regions and half-edge data structures was adapted for MATLAB.

Notes

• Console logs include execution times for key steps (e.g., point generation, contour extraction).
• Disable plotting (plot_* = false) to speed up execution if only data is needed.
• For large geometries, reduce d or iterations to optimize performance.

License

MIT License. See LICENSE file for details.

Author

Nicolás Jesús Muñoz Guamán (nicolasjmunoz@gmail.com / nmunoz@mechaide.com)

Source

Original code in JS: www.tesis.mechaide.com