Low-cost multi-view pose tracking using active markers
Authors:
- 1 May 2019
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 261-268
Abstract
Vision-based motion tracking systems have important applications in robotics and human motion analysis as well as in the entertainment industry. In this paper, we present an overview of a high-precision low-cost motion tracking system that combines active pulsed markers with multi-view pose reconstruction. While many passive-marker tracking systems impose restrictions on marker placement for correct marker identification, active pulsed LEDs allow for rigid body tracking with arbitrary marker placement as well as marker tracking on unknown articulated and deformable bodies. Individual markers are tracked via a path search method, which is robust against short temporary occlusions. Once detected, a rigid body can be tracked via a minimal number of marker observations from a single view or distributed across multiple views, even if the position of each marker could not be reconstructed. Additionally, IMU data can be fused with optical marker tracking to improve precision and robustness against marker occlusions.Keywords
This publication has 31 references indexed in Scilit:
- Vision-based hand pose estimation: A reviewComputer Vision and Image Understanding, 2007
- An Introduction to Inertial and Visual SensingThe International Journal of Robotics Research, 2007
- A survey of advances in vision-based human motion capture and analysisComputer Vision and Image Understanding, 2006
- Fusion of Vision and Inertial Data for Motion and Structure EstimationJournal of Robotic Systems, 2004
- Principles of Fusion of Inertial Navigation and Dynamic VisionJournal of Robotic Systems, 2003
- Inertial sensor technology trendsIEEE Sensors Journal, 2001
- A flexible new technique for camera calibrationIEEE Transactions on Pattern Analysis and Machine Intelligence, 2000
- Motion and structure from feature correspondences: a reviewProceedings of the IEEE, 1994
- Trinocular stereo vision for roboticsIEEE Transactions on Pattern Analysis and Machine Intelligence, 1991
- A computer algorithm for reconstructing a scene from two projectionsNature, 1981