Alexandre Bry

Study of a deep learning model for identifying individual trees from aerial photographs and LiDAR point clouds. To this end, I had to: - Assemble and create a dataset containing images (visible and infrared spectra), a point cloud and individual annotations for trees in Amsterdam - Modify and implement a deep learning model capable of taking images and the point cloud as input in order to extract information from both sources - Study the model's performance, particularly with regard to covered trees, i.e. trees that are partially or completely hidden by other trees seen from above. Report: https://alexandre-bry.github.io/Geodan-internship-report/

Improvement of the eco-design tools of Artelia's Environment department and its approach to linear infrastructure projects (tramways, buses, and roads): - Development of a modular and automated tool for eco design in a project - Literature review and drafting of fact sheets on materials of interest for different aspects of eco design