Façade segmentation from 3D Point Clouds for vehicle localization using Openstreetmap

The project involves the integration of 3D point clouds acquired using a LiDAR (Velodyne VLP-16) in the OpenStreetMap service. The activity is aimed at improving the location of the vehicle within a known topological map, taking advantage of the Road Layout Estimation framework.

The project is aimed at finding a way to include 3D information, acquired from a LIDAR scanner collected within the university campus through the use of LIDAR sensors mounted on the USAD vehicle, inside maps from Open Street Maps project. In the initial phase there is going to be a data. The student is expected to acquire the ability to understand how to insert the point cloud in the maps using appropriate simulation systems. The student is also expected to acquire the ability to understand the functioning of Road Layout Estimation and of the ROS framework, any programming activity will take place in C ++.

With this work we to experiment  the benefits of realistic, high definition point cloud in the work  proposed in "Leveraging the OSM building data to enhance the localization of an urban vehicle" and "Vehicle localization using 3D building models and Point Cloud matching"