Host institution: University of Cambridge
Project Title: Updating model geometry from registered point cloud and image datasets
Objectives: To develop an automatic model updating technique that registers the real-world data (point clouds and images) on the existing BIM model (i.e. a design model or a previous as-is model), matches real with virtual elements, detects missing or additional objects and relationships, and updates existing model geometrical properties and relationships to match those derived from the real-world data. The ESR must: (1) define the properties and relationships to be updated; (2) devise the technique; (3) implement and test a prototype; (4) drive the technology transfer process through development and IP exploitation with partner LocLab Consulting.
Expected Results: The ESR will develop a prototype implementation of the model updating technique capable of automatically updating the number, type, location, dimensions, relationships and textures of model objects. It will also borrow learning strategies explored by ESR 1 to insert additional real objects that are not present in the virtual model. To do so, the fellow will require extensive training on computer vision, machine learning and computational geometry (at STFD), modelling standards and geometry enrichment (Technion – Israel Institute of Technology). A secondment to LocLab Consulting is planned where the ESR will develop the potential for application of this technique to a real application scenario.
Enrolment in Doctoral degree: University of Cambridge, PhD Engineering.
Advisor: Ioannis Brilakis