<p>This paper presents the development of a novel dual active marker system designed to address the challenges of traditional marker-based pose estimation methods. Conventional approaches relying on multiple 2D or volumetric markers often face issues such as marker loss, physical damage, and limited recognition from specific angles, especially in large workspaces. To overcome these limitations, the proposed system integrates two synchronized marker tracking modules, each equipped with an embedded computer, a 2-axis actuator, an augmented reality (AR) marker, and a camera. These modules actively track each other’s markers, enabling reliable and robust 6D pose estimation. By aggregating tracking data from both modules and applying a Kalman filter, the system achieves high accuracy in estimating relative positions and orientations. Experimental results demonstrate that the dual active marker system not only enhances pose estimation precision but also provides scalability and reliability for applications in robotics, augmented reality, and virtual reality.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Robust 6D Pose Estimation Using Dual Active Marker System

  • Hyeon-Ju Choi,
  • Yeong-Bin Kim,
  • Bum Yong Park,
  • Dong-Hyun Lee

摘要

This paper presents the development of a novel dual active marker system designed to address the challenges of traditional marker-based pose estimation methods. Conventional approaches relying on multiple 2D or volumetric markers often face issues such as marker loss, physical damage, and limited recognition from specific angles, especially in large workspaces. To overcome these limitations, the proposed system integrates two synchronized marker tracking modules, each equipped with an embedded computer, a 2-axis actuator, an augmented reality (AR) marker, and a camera. These modules actively track each other’s markers, enabling reliable and robust 6D pose estimation. By aggregating tracking data from both modules and applying a Kalman filter, the system achieves high accuracy in estimating relative positions and orientations. Experimental results demonstrate that the dual active marker system not only enhances pose estimation precision but also provides scalability and reliability for applications in robotics, augmented reality, and virtual reality.