Clément Colin Ph.D. thesis defense : “Multi-scale geospatial data management and visualisation applied for CMMS”

Tuesday 18th of June at 10h a.m. Paris time, Clément Colin, Ph.D. Candidate has defended his thesis named “Multi-scale geospatial data management and visualisation applied for CMMS”. His thesis defense took place at the conference room of the Faculty “IUT Lumière Lyon 2” in Villeurbanne, France. Take a look at the summary below :

Summary :

Cities and the objects that make them up, such as buildings, water, electricity and road networks, have increasingly precise digital twins that play an important role in understanding territories. The growing use of Geographic Information Systems (GIS), Building Information Model (BIM) and City Information Model (CIM) has led to the creation of a large number of geospatial representations of these urban objects, made up of geometric and semantic data, structured by numerous standards. These representations provide a variety of thematic and spatial information to describe what these objects are physically, functionally and operationally. A better understanding of these urban objects can be provided by applications enabling users to access, visualize and analyze these urban objects using these different representations.

In this thesis, we focus on multiscalar interactive web navigation and visualization of multiple representations of the same object.  We will consider various heterogeneous standards for representing the interior and exterior of a building and a city. Our first two contributions enable the creation of navigable and contextual views of these heterogeneous representations in a single web context, using approaches based on data integration methods. To this end, we propose a methodology and an open-source tool, Py3DTilers, for extracting, manipulating and visualizing the geometry of geospatial data, as well as a model-based semantic data integration methodology, to ensure that all the information present in these data can be brought and understood by the users. Our third contribution is the formalization of the concepts of Variant -instance or set of instances representing the same entity- and Variant Identifier to reference and navigate through a set of representations of the same object. Finally, our last contribution focuses on the choice of geometric representation of an object to be displayed, depending on the users’ 3D context. We propose a study of the levels of detail described in different geospatial data standards, as well as a metric for describing the complexity of a geometric representation to enable this choice.

Finally, this thesis was carried out in partnership with Carl Software – Berger-Levrault, a publisher of computer-aided maintenance software and asset management solutions. Particular attention was paid to interoperability (use of standards), reusability (creation of shared software architecture based on open-source tools) and reproducibility of the proposed solutions. This thesis aims to improve the understanding of equipment to facilitate its maintenance and management, by allowing the 3D visualization of equipment and the exploitation of the knowledge that can be found in various representations. This is achieved by establishing a natural link between equipment representations existing in this domain and various geospatial data sources.

The jury was composed of:


  • M. Bouju Alain, Maître de conférences HDR, Université de la Rochelle
  • Mme. Morin Géraldine, Professeure, Institut National Polytechnique de Toulouse
  • Mme. Villanova-Oliver Marlène, Professeure, Université Grenobles Alpes
  • Mme. Puissant Anne, Professeure, Université de Strasbourg
  • M. Josselin Didier, Directeur de recherches, CNRS

Supervisors :

  • M. Gesquiere Gilles, Professeur, Université Lumière Lyon 2
  • M. Samuel John, Enseignant chercheur, CPE Lyon
  • Mme. Servigne Sylvie, Maîtresse de conférences, INSA Lyon
  • M. Christophe Bortolaso, Responsable de la recherche Appliquée, Berger-Levrault

Want to watch the video about his thesis work ? Click below

More ...

Scroll to Top