EDEN / Rovers Navigation / Perception / Panoramic Vision
 [ People Involved ] [ Related Publications ]

Seeing all around

Vision is one of the richest source of information in robotics. However, conventional cameras have a rather limited field of view, which raises various problems (and in particular the necessity to control their orientation). Panoramic imagers solve this problem, most of the time by using a convex mirror. We choose to equip Dala with such a sensor, which we bought to the remote reality company. Some images and a small video can be seen in the gallery. This sensor has the great advantage to have a single center of projection, as any desired image projected on any image plane can be generated, such as a pure perspective image for instance. We currently use the produced images to qualitatively localize the rover.

Also, we recently initiated some work on panoramic stereovision [Gonzalez-Barbosa 2003].

The sensor mounted on Dala

Principle of image formation



Related Links

Only one is necessary here : the nice Omnidirectional vision page set up by Kostas Daniilidis, that gathers a set of links to various universities, labs and companies involved in the development and exploitation of panoramic cameras..

Related Publications

[Gonzalez-Barbosa 2003]  [related pages] [abstract] [download] [BibTeX]  [top]

J-J. Gonzalez-Barbosa and S. Lacroix. Un algorithme rapide de stéréovision panoramique dense. Technical report, LAAS/CNRS. (In french) 2003.




People Involved

Simon Lacroix, José Gonzalez.
  General Information
 About the EDEN Project
 Photo Gallery
*  Panoramic images
 Videos
 People
 Publications
 Related Projects
 Job opportunities
 Intranet

  Robots
 Adam
 Lama
 The Blimp
*Dala

  Rovers Navigation
 Motion control
 Environment Perception
   Stereo-vision
   Image segmentation
>  Panoramic vision
   Interest points matching
   Pixel tracking
 Environment Modeling
 Localization
*  Panoramic images
 Motion Generation
 Integration

  Autonomous Blimps
 Control
 Environment Perception

  Multi-Robot Cooperation
 Exploration