The constrained SLAM framework for non-instrumented augmented reality - CEA - Commissariat à l’énergie atomique et aux énergies alternatives Access content directly
Journal Articles Multimedia Tools and Applications Year : 2016

The constrained SLAM framework for non-instrumented augmented reality


This paper addresses the challenging issue of marker less tracking for Augmented Reality. It proposes a real-time camera localization in a partially known environment, i.e. for which a geometric 3D model of one static object in the scene is available. We propose to take benefit from this geometric model to improve the localization of keyframe-based SLAM by constraining the local bundle adjustment process with this additional information. We demonstrate the advantages of this solution, called contrained SLAM, on both synthetic and real data and present very convincing augmentation of 3D objects in real-time. Using this tracker, we also propose an interactive augmented reality system for training application. This system, based on a Optical See-Through Head Mounted Display, allows to augment the users vision field with virtual information accurately co-registered with the real world. To keep greatly benefit of the potential of this hand free device, the system combines the tracker module with a simple user-interaction vision-based module to provide overlaid information in response to user requests.
Not file

Dates and versions

cea-01830529 , version 1 (05-07-2018)



M. Tamaazousti, S. Naudet-Collette, V. Gay-Bellile, S. Bourgeois, B. Besbes, et al.. The constrained SLAM framework for non-instrumented augmented reality. Multimedia Tools and Applications, 2016, 75 (16), pp.9511 - 9547. ⟨10.1007/s11042-015-2968-8⟩. ⟨cea-01830529⟩
60 View
0 Download



Gmail Facebook Twitter LinkedIn More