Scholarlink

<p>Fiducial markers are fundamental tools for visual localization, calibration, and augmented reality; however, the precision of these systems is fundamentally constrained by the spatial configuration of the markers. We introduce a principled pipeline for optimal marker deployment in indoor scenes. The method (i) extracts wall-like candidate locations from dense 3D reconstructions, (ii) computes a Fisher Information Matrix (FIM) for each candidate by analytically aggregating the Jacobians of projected marker corners over a set of camera poses, and (iii) selects a compact subset of markers via an information-aware optimization strategy. Given a Matterport-style point cloud, the system automatically proposes marker poses on planar surfaces at user-specified heights. The proposed framework was validated through both simulation and real-world indoor experiments and compared against random, uniform, and optimization-based baseline strategies.</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

Visibility-aware Fisher information optimization for optimal fiducial marker deployment in complex indoor environments

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月19日

查看全文
  • Banafshe Akbarinia,
  • Meor Faisal Zulkifli,
  • Bushroa Abd Razak,
  • Mohd Ridha Muhamad,
  • Hamed Shahmohamadi Ousaloo,
  • Norrima Mokhtar

摘要

Fiducial markers are fundamental tools for visual localization, calibration, and augmented reality; however, the precision of these systems is fundamentally constrained by the spatial configuration of the markers. We introduce a principled pipeline for optimal marker deployment in indoor scenes. The method (i) extracts wall-like candidate locations from dense 3D reconstructions, (ii) computes a Fisher Information Matrix (FIM) for each candidate by analytically aggregating the Jacobians of projected marker corners over a set of camera poses, and (iii) selects a compact subset of markers via an information-aware optimization strategy. Given a Matterport-style point cloud, the system automatically proposes marker poses on planar surfaces at user-specified heights. The proposed framework was validated through both simulation and real-world indoor experiments and compared against random, uniform, and optimization-based baseline strategies.