Scholarlink

<p>This study aims to resolve the inherent size ambiguity in monocular visual simultaneous localization and mapping (VSLAM) systems while minimizing sensor and computational requirements. We address the research question: <i>can a low-cost laser beam provide sufficient spatial constraints to restore absolute size and enable efficient 6-DoF VSLAM without multi-sensor setups?</i> We propose a lightweight framework integrating a single camera with a sub-$1 laser beam emitter. The laser projects a spatial constraint prior, providing absolute depth references during motion estimation. By embedding this depth data into the visual SLAM pipeline, we establish size-aware optimization without external sensors like IMUs or depth cameras. The system eliminates complex calibration and operates with minimal computational overhead. Hardware experiments and dataset evaluations demonstrate successful absolute size recovery in resource-constrained platforms. Compared to ORB-SLAM3 and VINS-Mono, our method reduces computational power consumption by 27% while maintaining comparable pose estimation precision. The total hardware cost remains under $5, requiring a small amount of memory for real-time operation on embedded platforms. This work presents the first 6-DoF VSLAM system achieving metric-size awareness through ultra-low-cost laser constraints. By synergizing geometric priors with visual odometry, we overcome size drift while preserving computational efficiency. The low cost and lightweight performance make this solution particularly viable for micro-robots, front-end devices, and other resource-constrained applications where traditional multi-sensor approaches are impractical.</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

Ultra-lightweight 6-DoF VSLAM with absolute size restoration using low-cost laser beam priors

  • 期刊论文

  • 发表日期：

    2026年3月10日

查看全文
  • Hao Xu,
  • Liping Li,
  • Xixiang Liu

摘要

This study aims to resolve the inherent size ambiguity in monocular visual simultaneous localization and mapping (VSLAM) systems while minimizing sensor and computational requirements. We address the research question: can a low-cost laser beam provide sufficient spatial constraints to restore absolute size and enable efficient 6-DoF VSLAM without multi-sensor setups? We propose a lightweight framework integrating a single camera with a sub-$1 laser beam emitter. The laser projects a spatial constraint prior, providing absolute depth references during motion estimation. By embedding this depth data into the visual SLAM pipeline, we establish size-aware optimization without external sensors like IMUs or depth cameras. The system eliminates complex calibration and operates with minimal computational overhead. Hardware experiments and dataset evaluations demonstrate successful absolute size recovery in resource-constrained platforms. Compared to ORB-SLAM3 and VINS-Mono, our method reduces computational power consumption by 27% while maintaining comparable pose estimation precision. The total hardware cost remains under $5, requiring a small amount of memory for real-time operation on embedded platforms. This work presents the first 6-DoF VSLAM system achieving metric-size awareness through ultra-low-cost laser constraints. By synergizing geometric priors with visual odometry, we overcome size drift while preserving computational efficiency. The low cost and lightweight performance make this solution particularly viable for micro-robots, front-end devices, and other resource-constrained applications where traditional multi-sensor approaches are impractical.