Generating LOD-1 representation of buildings using CGAL

[DorianG02]

Internship on generating LOD-1 representation of buildings using CGAL

• Intern : Dorian Geraldes Pereira @DorianG02 (Master 1 CSMI - https://csmi.cemosis.fr)
• Advisors : @vincentchabannes and @palliez

Overview

This project is part of the ExaMA and Hidalgo2 initiative, where Cemosis aims to implement new tools to improve Urban Energy Simulation. Our work focuses on generating 3D models of buildings. The goal of this internship is to develop a geometric reconstruction tool for LOD1 buildings. This level considers buildings as collections of polygons extruded by height, with specifications for roof shapes. Based on this information, we aim to reconstruct a 3D mesh that adheres to geometric conformity constraints.

The main objectives of this project include creating accurate and detailed surface meshes and 3D models that can be used for energy simulations, urban planning, and other applications. The tool developed during this internship will enhance the ability to analyze and simulate urban environments efficiently.

Objectives

These are the objectives of this internship:

• Generation of a surface mesh representing the building’s external envelope.
  This involves creating a detailed and accurate mesh that outlines the outer surface of the building, ensuring that all exterior features are captured correctly.

• Addition of internal walls delimiting exterior walls and any slabs for each floor.
  Internal walls are added to define the layout of each floor, including any slabs that separate different levels within the building. This step is crucial for accurate floor planning and structural integrity.

• Roof modeling.
  The roof is modeled separately to ensure it accurately fits the building’s design. This involves creating a roof structure that aligns with the building’s overall architecture and specifications.

• Creation of a 3D building volume mesh (structure and/or indoor air).
  This objective focuses on generating a volumetric mesh that represents the building’s internal and external structures. This mesh is used for various simulations, including structural analysis and airflow studies.

• Mesh adaptation: quality control of the produced meshes.
  Ensuring the quality of the meshes is critical. This involves checking and refining the meshes to meet specific quality standards, making sure they are suitable for further analysis and use.

• Combining different parts of a building to accurately represent more complex shapes. For buildings with complex architectures, different parts need to be combined seamlessly. This objective ensures that all components fit together correctly to form a complete and accurate representation of the building.

Methodology

Here is the methodology I followed for this internship:

• Create the surface mesh of one building

  • Create the base: This step includes the triangulation of the polygon, the extrusion, and the creation of the internal wall.
  • Add the floor: This step includes the creation of the floor mesh and its merging with the base.
  • Add the roof: In this step, the roof is created using the input polygon with the Straight Skeleton algorithm from CGAL and is then merged with the base.

• Creation of the 3D mesh using the surface mesh

  • Convert the surface mesh: The surface mesh is converted in polyhedron of the cgal library to be usable in the 3D Mesh Generation algorithm
  • Create the 3D mesh: The polyhedron is used to create the 3D Mesh

• Expand the new functionnality to create multiple building: The algorithm to create the surface and 3D mesh is now able to generate multiple building in a single mesh

Result

Surface Mesh

In this section the different color represent a marker

Here a close visualisation of 2 surface mesh:

Here a visualisation of a district of Lingolsheim:

Volume Mesh

Here the visualisation of a 3D volume mesh:

To find out more

More information are available at ktirio geom doc

References

• The CGAL Project. CGAL User and Reference Manual. CGAL Editorial Board, 6.0 edition, 2024. link

• Pierre Alliez, Clément Jamin, Laurent Rineau, Stéphane Tayeb, Jane Tournois, and Mariette Yvinec. 3D Mesh Generation. In CGAL User and Reference Manual. CGAL Editorial Board, 6.0 edition, 2024. link

• Fernando Cacciola, Sébastien Loriot, and Mael Rouxel-Labbé. 2D Straight Skeleton and Polygon Offsetting. In CGAL User and Reference Manual. CGAL Editorial Board, 6.0 edition, 2024. link

• David Coeurjolly, Jacques-Olivier Lachaud, Konstantinos Katrioplas, Sébastien Loriot, Ivan Pađen, Mael Rouxel-Labbé, Hossam Saeed, Jane Tournois, and Ilker O. Yaz. Polygon Mesh Processing. In CGAL User and Reference Manual. CGAL Editorial Board, 6.0 edition, 2024. link

• Feelpp. Open-Source C++ library. link

• Cemosis. Innovation through modeling, simulation, optimization, and high performance computing. link

• Irma. Institute for Advanced Mathematical Research. link

• Numpex. French program dedicated to Exascale. link

• Hidalgo2. HPC and big data technologies for global challenges. link

• Meshlab. Open source system for processing and editing 3D triangular meshes. link

• Paraview. Open source post-processing visualization engine. link

• UFR. link

• STL, File Format. link

• JSON, File Format. link

• VTU, File Format derived from VTK. link