| 日時(Date) | 平成30年11月27日（火）2限（11:00--12:30) Tue. Nov. 27th, 2018, 2nd Period (11:00--12:30) |
|---|---|
| 場所(Location) | L1 |
| 司会(Chair) | 舩冨卓哉 准教授 |
| 講演者(Presenter) | Cedric Demonceaux (Prof. Universite de Bourgogne 4. ) Guillaume Caron (Assoc. Prof. Universite de Picardie Jules Verne 4.) Nicolas Ragot (Researcher, ESIGELEC–IRSEEM) |
| 題目(Title) | Cedric Demonceaux (3D reconstruction and localization in a dynamic world) Guillaume Caron (Reliable planar object pose estimation in light-fields from best sub-aperture camera pairs) Nicolas Ragot (Omnidirectional catadioptric vision - modelling and processing) |
| 概要(Abstract) | Cedric Demonceaux (Nowadays, it becomes more and more simple to reconstruction in 3D a scene using a single camera with structure from motion techniques, RGB-D cameras or lidar. Nevertheless, when the scene is composed of many dynamic objects, the main approaches do not work correctly. In this talk, we will discuss how we can analyze with a RGB-D sensors a 3D scene including moving objects in order to reconstruct separately all the objects and retain only the fixed part that can be used to localize a robot.) Guillaume Caron (A light-field camera can obtain richer information about a scene than a usual camera. This property offers a lot of potential for robot vision. In this talk, we present a method for pose estimation of a planar object with a light-field camera. The light-field camera can be regarded as a set of sub-aperture cameras. Although any combination of them can theoretically be used for the pose estimation, the accuracy depends on the combination. We show that the estimated pose error can be reduced by selecting the best pair of sub-aperture cameras. We have evaluated the accuracy of our approach with real experiments using a light-field camera in front of planar targets held by an industrial manipulator for ground truth comparison. ) Nicolas Ragot (Since the late 1990s, works have been dedicated to design and evaluate vision systems that allow an observation of a scene with a large field of view, panoramic or omnidirectional. These systems have proven to be particularly suitable for robotic applications as they provide a view of the environment independent from a specific direction and without any temporal failure. Several solutions can be used to increase the field of view of a camera: a camera performing a rotation, several points of view from a camera network. In this talk, we will focus on catadioptric omnidirectional vision systems that combine a camera and a convex mirror. This presentation will be an opportunity to present this sensor technology, their modeling (ad-hoc or from the unified model) and to illustrate their uses through some examples.) |
| 講演言語(Language) | English |
| 講演者紹介(Introduction of Lecturer) | Cedric Demonceaux (received the M.S. degree in Mathematics in 2001 and the PhD degree in Image Processing from the Universite de Picardie Jules Verne in 2004. In 2005, he became associate professor at MIS-UPJV. From 2010 to 2014, he has been an CNRS-Higher Education chair at Le2I UMR CNRS, Universite de Bourgogne. Since 2014, he is full Professor at the University of Burgundy. His research interests are in image processing, computer vision and robotics.) Guillaume Caron (is Associate Professor since 2011 and he is heading the Robotic Perception group of the MIS laboratory since 2016 at Universite de Picardie Jules Verne. He received the Ph.D. degree in robotics from the same university in 2010. He spent one year as a postdoctoral associate at INRIA Rennes. He was also a visiting research scholar at the University of Osaka for two months in 2013. His research interests include artificial vision for robotics, real-time visual tracking and servoing. ) Nicolas Ragot (received a Master of Engineering from ESIGELEC in 2002 and the Master of Science degree in Electrical Engineering from Université Paris XI in 2003. He earned his Ph.D. in Instrumentation and Vision Sensor Control from Université de Rouen in 2009. Since 2009, he is a lecturer-researcher in the Instrumentation, IT and Systems Lab at ESIGELEC–IRSEEM. His research interests deal with computer vision, omnidirectional vision and their applications to perception for intelligent vehicles.) |