Closing the Loop in Appearance-Guided Structure-from-Motion for Omnidirectional Cameras

Abstract : In this paper, we present a method that allows us to recover a 400 meter tra jectory purely from monocular omnidirectional images very accurately. The method uses a novel combination of appearance-guided structure from motion and loop closing. The appearance-guided monocular structure-from-motion scheme is used for initial motion estimation. Appearance information is used to correct the rotation estimates computed from feature points only. A place recognition scheme is employed for loop detection, which works with a visual word based approach. Loop closing is done by bundle adjustment minimizing the reprojection error of feature matches. The proposed method is successfully demonstrated on videos from an automotive platform. The experiments show that the use of appearance information leads to superior motion estimates compared to a purely feature based approach. And we demonstrate a working loop closing method which eliminates the residual drift errors of the motion estimation. Note that the recovered tra jectory is one of the longest ones ever reported with a single omnidirectional camera.
Type de document :
Communication dans un congrès
The 8th Workshop on Omnidirectional Vision, Camera Networks and Non-classical Cameras - OMNIVIS, Oct 2008, Marseille, France. 2008
Liste complète des métadonnées

Littérature citée [22 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/inria-00325324
Contributeur : Peter Sturm <>
Soumis le : dimanche 28 septembre 2008 - 22:00:29
Dernière modification le : lundi 29 septembre 2008 - 09:19:12
Document(s) archivé(s) le : vendredi 4 juin 2010 - 11:52:56

Fichier

A14CR.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : inria-00325324, version 1

Collections

Citation

Davide Scaramuzza, Friedrich Fraundorfer, Marc Pollefeys, Roland Siegwart. Closing the Loop in Appearance-Guided Structure-from-Motion for Omnidirectional Cameras. The 8th Workshop on Omnidirectional Vision, Camera Networks and Non-classical Cameras - OMNIVIS, Oct 2008, Marseille, France. 2008. 〈inria-00325324〉

Partager

Métriques

Consultations de la notice

210

Téléchargements de fichiers

189