Editorial
Omnidirectional robot vision

https://doi.org/10.1016/j.robot.2010.02.006Get rights and content

Introduction

Omnidirectional vision research has always been greatly stimulated by mobile robotics and it has been finding there its major applications. Significant progresses have been made since the pioneer works, in which shielded light bulbs or steel spheres were used as mirrors placed above the robots’ cameras. These cameras were strangely pointed upward, instead of forward, as in all other robots at that times. Recently, a broader community has started working with omnidirectional sensors and commercial applications appeared also in surveillance and automotive industry. Several companies are already selling high quality omnidirectional mirrors and even complete omnidirectional cameras.

The increase of resolution of modern cameras overcomes one of the major limitations of omnidirectional vision: the low resolution of omnidirectional images. At the same time, the increase of computational power on board of nowadays robots enables the use of more complex robot vision algorithms. Therefore, more and more computational intensive computer vision algorithms are now applicable to robot vision. In the past few years, several robots mounting one or more omnidirectional cameras have been created. We have seen wheeled robot, flying robots, even humanoid robots mounting omnidirectional cameras.

Despite this, the robotics community has not been very active in disseminating its results to the scientific community. On the contrary, the computer vision community has been much more active. Yearly, the Workshop on Omnidirectional Vision, Camera Networks and Non-classical Cameras has been organized in different Computer Vision conferences, but it was organized just once in a robotics conferences. This is reflected by the fact that most of the papers in the proceedings of the Omnidirectional Vision Workshop are coming from the computer vision community. We believe that to foster dissemination from the robotics community and to create active collaboration between researchers in computer vision and robotics, we need to have an omnidirectional vision workshop also in a robotics conference. The first edition of the Omnidirectional Robot Vision workshop was held in Venice (Italy) in Nov. 2008. This was co-located with the conference on Simulation, Modeling and Programming for Autonomous Robots (SIMPAR 2008). The workshop had a large success of submissions (19 papers were submitted) and of attendees (more than 30 people attended the workshop). After the workshop, we promoted a special issue of the journal Robotics and Autonomous Systems opened not only to papers presented at the workshop, but also to non-previously published papers. We received a large number of papers and the review process was very hard. The current special issue contains the 7 best papers selected among the 18 papers submitted. We think the great respond of the robotics community, both to the workshop and to the journal special issue, confirmed we were right in identifying the lack of an appropriate place for dissemination of the scientific results of omnidirectional vision applied to robots.

Section snippets

This issue

We have divided this special issue in two parts. The first part consists of five papers proposing new tools and algorithms for omnidirectional vision applied to wheeled mobile robots. The second part consists of two papers and illustrates applications of omnidirectional vision to other domains like automotive and avionic industries.

In the first part, the first paper of Lui and Jarvis presents a new omnidirectional stereo vision system based on two catadioptric omnidirectional cameras, in which

Acknowledgements

We wish to thank all the authors and reviewers of the papers in this special issue for their great work in improving, polishing, and commenting the papers you can finally find in the next pages. We also wish to thank Prof. Rüdiger Dillmann, the Editor in Chief of Robotics and Autonomous System, for providing us the opportunity to publish this special issue, and Prof. Enrico Pagello, General Chair of the 2008 International Conference on Simulation, Modeling, and Programming for Autonomous Robots

References (0)

View full text