Wide-area scene mapping for mobile visual tracking

Ventura, Jonathan; Höllerer, Tobias

doi:10.1109/ismar.2012.6402531

Cited by 55 publications

(23 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The matching is distributed over several frames to keep processing times low. Ventura and Höllerer [29] propose to query an image-based localization server to estimate the camera pose of the first image, allowing subsequent real-time pose tracking relative to the server-side 3D model, which is also kept locally on the mobile device. To efficiently match features in the following images to this model, the pose of the previous image is used to cull 3D points behind the camera and outside the image.…”

Section: Overall Approachmentioning

confidence: 99%

“…Nevertheless, the flexibility of mobile devices creates a strong need for largescale mobile localization, e.g., for city-scale augmented reality. Recently, Lim et al [22] and Ventura and Höllerer [29] proposed approaches for real-time, largescale mobile pose tracking. However, both approaches require to keep a 3D model of the environment on the device, which can already consume more than 100MB for a scene of size 8m × 5m [22], limiting their applicability for mobile devices with their hard memory restrictions.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Scalable 6-DOF Localization on Mobile Devices

Middelberg

Sattler

Untzelmann

et al. 2014

Lecture Notes in Computer Science

170

147

View full text Add to dashboard Cite

Abstract. Recent improvements in image-based localization have produced powerful methods that scale up to the massive 3D models emerging from modern Structure-from-Motion techniques. However, these approaches are too resource intensive to run in real-time, let alone to be implemented on mobile devices. In this paper, we propose to combine the scalability of such a global localization system running on a server with the speed and precision of a local pose tracker on a mobile device. Our approach is both scalable and drift-free by design and eliminates the need for loop closure. We propose two strategies to combine the information provided by local tracking and global localization. We evaluate our system on a large-scale dataset of the historic inner city of Aachen where it achieves interactive framerates at a localization error of less than 50cm while using less than 5MB of memory on the mobile device.

show abstract

Section: Overall Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Scalable 6-DOF Localization on Mobile Devices

Middelberg

Sattler

Untzelmann

et al. 2014

Lecture Notes in Computer Science

170

147

View full text Add to dashboard Cite

show abstract

“…More scalable approaches assume the availability of an external server, which sends the relevant model parts to the device [4] or even performs the actual localization [32,51,52]. The latter methods run SLAM on the device, enabling them to handle the latency of transmitting images to the server and to track the pose for periods where localization against the global model fails [32,44,52].…”

Section: Introduction and Related Workmentioning

confidence: 99%

Get Out of My Lab: Large-scale, Real-Time Visual-Inertial Localization

Lynen

Sattler

Bosse³

et al. 2015

Robotics: Science and Systems XI

243

210

View full text Add to dashboard Cite

Abstract-Accurately estimating a robot's pose relative to a global scene model and precisely tracking the pose in real-time is a fundamental problem for navigation and obstacle avoidance tasks. Due to the computational complexity of localization against a large map and the memory consumed by the model, state-ofthe-art approaches are either limited to small workspaces or rely on a server-side system to query the global model while tracking the pose locally. The latter approaches face the problem of smoothly integrating the server's pose estimates into the trajectory computed locally to avoid temporal discontinuities.In this paper, we demonstrate that large-scale, real-time pose estimation and tracking can be performed on mobile platforms with limited resources without the use of an external server. This is achieved by employing map and descriptor compression schemes as well as efficient search algorithms from computer vision. We derive a formulation for integrating the global pose information into a local state estimator that produces much smoother trajectories than current approaches. Through detailed experiments, we evaluate each of our design choices individually and document its impact on the overall system performance, demonstrating that our approach outperforms state-of-the-art algorithms for localization at scale.

show abstract

“…Since potential inaccuracies of the individual sensors add up, the positioning accuracy of this location-based approach is usually lower. There are some current research efforts such as the one presented by Ventura and Höllerer [16] that improve the positioning accuracy, but they often require explicit user input, special sensors or additional information.…”

Section: Related Workmentioning

confidence: 99%

Usability in 4D AR: Visualising Multi-temporal Real-time Geo-data in Augmented Reality Environments

Resch

Wichmann

Göll³

2015

Int. J. Interact. Mob. Technol.

View full text Add to dashboard Cite

Abstract-Even though advantages of 3D visualisation of multi-temporal geo-data versus 2D approaches have been widely proven, the particular pertaining challenge of realtime visualisation of geo-data in mobile Digital Earth applications has not been thoroughly tackled so far. In the emerging field of Augmented Reality (AR), research needs comprise finding the optimal information density, the interplay between orientation data in the background and other information layers, using the appropriate graphical variables for display, or selecting real-time base data with adequate quality and suitable spatial accuracy. In this paper we present a concept for integrating real-time data into 4D (three spatial dimensions plus time) AR environments, i.e., data with "high" spatial and temporal variations. We focus on three research challenges: 1.) high-performance integration of real-time data into AR; 2.) usability design in terms of displaying spatio-temporal developments and the interaction with the application; and 3.) design considerations regarding reality vs. virtuality, visualisation complexity and information density. We validated our approach in a prototypical application and extracted several limitations and future research areas including natural feature recognition, the cross-connection of (oftentimes monolithic) AR interface developments and well-established cartographic principles, or fostering the understanding of the temporal context in dynamic 4D Augmented Reality environments.

show abstract

Wide-area scene mapping for mobile visual tracking

Cited by 55 publications

References 25 publications

Scalable 6-DOF Localization on Mobile Devices

Scalable 6-DOF Localization on Mobile Devices

Get Out of My Lab: Large-scale, Real-Time Visual-Inertial Localization

Usability in 4D AR: Visualising Multi-temporal Real-time Geo-data in Augmented Reality Environments

Contact Info

Product

Resources

About