<p>Spherical markers are important in near-infrared (NIR) locators of orthopedic surgical robots. The spherical markers suffer from poor occlusion resistance due to occlusion by medical staff or blood contamination. Solutions to this challenge, such as adding markers or cameras, are limited by space and system complexity in the surgical environment. Compared with spherical markers, planar markers have better potential to enhance occlusion resistance due to their rich feature points. Current planar markers, which are designed for visible light locators in augmented reality (AR) applications, do not fulfill the dual requirements of occlusion resistance and miniaturization necessary for orthopedic surgical robots. Therefore, this paper will concentrate on the anti-occlusion of NIR small-size planar markers. The designed planar marker employs a checkerboard layout embedded in circular patterns to achieve orientation encoding, ID encoding, and redundant encoding mechanisms at the same time. A projective transformation with minimal representations of uncertain image points enhanced planar markers’ pose estimation. In translation and rotation experiments, the marker presented better results compared with two other planar marker systems: the classical ArUco and the advanced TopoTag. Moreover, the designed planar marker with redundant encoding can maintain pose estimation even with occlusion levels reaching up to 50%. Therefore, this paper offers a promising solution for enhancing the occlusion resistance of NIR planar markers.</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 NIR planar marker with occlusion resistance

  • Yunxiao Chen,
  • Weiqiang Zhang,
  • Fan Yang,
  • Ruihong Li,
  • Dezhao Lin,
  • Hongwei Chen

摘要

Spherical markers are important in near-infrared (NIR) locators of orthopedic surgical robots. The spherical markers suffer from poor occlusion resistance due to occlusion by medical staff or blood contamination. Solutions to this challenge, such as adding markers or cameras, are limited by space and system complexity in the surgical environment. Compared with spherical markers, planar markers have better potential to enhance occlusion resistance due to their rich feature points. Current planar markers, which are designed for visible light locators in augmented reality (AR) applications, do not fulfill the dual requirements of occlusion resistance and miniaturization necessary for orthopedic surgical robots. Therefore, this paper will concentrate on the anti-occlusion of NIR small-size planar markers. The designed planar marker employs a checkerboard layout embedded in circular patterns to achieve orientation encoding, ID encoding, and redundant encoding mechanisms at the same time. A projective transformation with minimal representations of uncertain image points enhanced planar markers’ pose estimation. In translation and rotation experiments, the marker presented better results compared with two other planar marker systems: the classical ArUco and the advanced TopoTag. Moreover, the designed planar marker with redundant encoding can maintain pose estimation even with occlusion levels reaching up to 50%. Therefore, this paper offers a promising solution for enhancing the occlusion resistance of NIR planar markers.