Séminaire (organisé par l’équipe de recherche ASER)

Mardi 4 décembre à 14h00 en salle Master

Yufeng YU

Université de Pékin

Visual Odometry assisted Vehicle localization in Structured-road environment

Résumé :

As we all know, normal GPS is not suitable for a high-accuracy localization. So camera based algorithms are widely used. For a dynamic environment localization, general visual odometry witch based on feature points cannot give a good result. We humans localize ourselves based on not only eyes, but also knowledge. Visual odometry needs knowledge or assumptions too. One quite simple assumption for on-road environment is that the road is flat and the crosses are right angle bended (see Fig. 1). Using this assumption, the rotation parameters can be settled easily based on line segments without accumulative error. Then the position can be solved using normal odometry method. If we have a sketch map, right angle assumption is not necessary.

Mengwen HE

Université de Pékin

Multi-LIDAR calibration and/or Velodyn-based 3D SLAM

Résumé :

Mobile robots equipped with cheap single-beam LIDARs are well suited for acquiring 3D data of environment. But accurate calibration for data fusion is required. Existing calibration methods for multi-LIDAR require specific calibration target or known environment. We intend to develop a calibration method which requires no specifically placed calibration targets, no special environments, no or few manual operations on software. Our approach uses the multi-type geometry features in outdoor point cloud, such as point, line, plane or quadric. In order to increase its robustness and accuracy, we defined a probabilistic distance measure to combine all types of geometry features into one registration process in same scale.


Mardi 24 octobre 2017

Séminaire à 14h00 en GI042 (Bâtiment Blaise Pascal - UTC) présenté par Richard NELSON, Maître de conférences à l’Université de Waikato à Hamilton (Nouvelle-Zélande).
WAND : Network Research at the University of Waikato.

Mardi 17 octobre 2017

Séminaire à 14 h en GI042 (Bâtiment Blaise Pascal), présenté par Corina Sandu, Professeur à Virginia Tech. « Overview of Dr. Corina Sandu’s Research and Teaching at Virginia Tech ».

Mardi 4 avril 2017

Séminaire à 14 h dans l’amphi du Centre d’Innovation de l’UTC, présenté par Xavier LAGORCE, PhD, Head of Computer Vision, Chronocam.
« Chronocam : Event-based cameras for machine vision »

Mardi 22 novembre 2016

Séminaire à 14 h au Centre d’Innovation de l’UTC, présenté par Riccardo SPICA, Lagadic group, Inria Rennes Bretagne Atlantique & IRISA.
« Active visual estimation for single and multiple robot systems »

Pages 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | ...


FR SHIC 3272

Collegium UTC/CNRS