CADART, M Nicolas (2019) Perception de la route par traitement de nuages de points LiDAR pour le véhicule autonome PFE - Project Graduation, ENSTA.



Road detection is a key issue to ensure a robust navigation for autonomous vehicles. Main sensor in such perception systems, the Velodyne LiDAR provides high resolution and omnidirectional information of the car's surroundings. Therefore, this project adresses the processing of 3D point clouds to produce fast and reliable segmentation and mapping of the navigable space. Named RoadSeg, the proposed solution uses a specially designed light-weight convolutional neural network to predict pixel-wise semantic segmentation from spherical LiDAR images. Thanks to an evidential framework, a fusion scheme aggregates labelled scans and produces dense local maps as 2D occupancy grids intended for navigation. Experiments conducted on an on purpose collected dataset as well as field tests show great performances while ensuring real-time processing.

Item Type:Thesis (PFE - Project Graduation)
Uncontrolled Keywords:autonomous vehicle, robotics, ROS, deep learning, CNN, FCN, Velodyne LiDAR, point cloud processing, semantic segmentation, obstacles detection, evidential grids, occupancy grids, visual odometry
Subjects:Information and Communication Sciences and Technologies
Mathematics and Applications
ID Code:7509
Deposited By:Nicolas Cadart
Deposited On:27 janv. 2020 15:19
Dernière modification:27 janv. 2020 15:19

Repository Staff Only: item control page