Publication type: Conference other
Type of review: No review
Title: Real time motion tracking for augmented reality with TOF camera and vulkan rendering
Authors: Wegmann, Marcel
et. al: No
Conference details: Embedded Computing Conference (ECC), Winterthur, 31. Mai 2022
Issue Date: 31-May-2022
Language: English
Subjects: Augmented reality; Time of Flight Camera; TOF; AR; RANSAC; Singular Value Decomposition (SVD); Kalman Filter; Sensor Fusion; Vulkan API
Subject (DDC): 006: Special computer methods
Abstract: This presentation describes a new concept for motion tracking in augmented reality systems using new approaches from an actual research project at ZHAW-InES. Motion tracking and the resulting recalculation of the representation of virtual objects is the key function in augmented reality systems. One of the innovations described in this paper is the sensor fusion of a Time of Flight (TOF) camera to spatially capture the environment in conjunction with a standard color camera to compute the 3D representation. The final 3D output is then generated using the open standard Vulkan low-overhead, cross-platform rendering API. Motion processing is based on the angular velocity and translation extraction using the Nvidia CUDA SIFT (Scale invariant feature transform) library. The extracted angular velocity is then fused with IMU (Inertial Motion Unit) data using a Kalman filter for the sensor fusion. The presentation also briefly discusses the hardware architecture, which is based on the NVIDIA Jetson-AGX. As a conclusion, motion tracking using a TOF camera and the Vulkan rendering library speeds up the tracking process, resulting in a real-time and more realistic user experience.
Further description: Vortrag in der Themengruppe Hardware und DSP
Fulltext version: Published version
License (according to publishing contract): Not specified
Departement: School of Engineering
Organisational Unit: Institute of Embedded Systems (InES)
Appears in collections:Publikationen School of Engineering

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.