The goal of this project is to design and develop efficient algorithms that produce in real-time a 3D model of the scene around a car from cameras mounted on the vehicle.
Simultaneous Localization and Mapping (SLAM) is concerned with the problem of building a map of an unknown environment by cameras or a mobile robot while at the same time navigating the environment using the map. The goal of this project is to design and develop efficient algorithms that produce in real-time a 3D model of the scene around a car from cameras mounted on the vehicle. In addition to performing SLAM, the developed algorithms will estimate a semantic class label for each reconstructed 3D point. Altogether, this will allow us to provide an accurate 3D map of the car’s surroundings.
The project will focus on two main applications. The first involves a parking scenario, where the car can be expected to move slowly and the environment to be mostly static. The second involves actual driving either in the city or on the highway.
This project lasts 3 years and is carried out in the context of the International Research Chair Drive for All led by MINES ParisTech (Prof. Arnaud de La Fortelle), sponsored by PSA, Safran and Valeo. It is led by the Computer Vision Laboratory of Prof Pascal Fua.
| Principal investigator | Prof. Pascal Fua |
| Project manager | Mathieu Salzmann |
| Sponsor | MINES ParisTech Foundation (Safran, Groupe PSA, Valeo) |
| Period | 2017-2019 |
| Laboratory | CVLAB |
| Collaboration | TRACE |