Scholarlink

<p>Three-dimensional (3D) imaging plays a crucial role in autonomous driving, medical diagnostics, and industrial inspection by providing comprehensive spatial information. Metalens-based 3D imaging is highly valued for imaging applications thanks to its compactness, with enhanced precision remaining a key research pursuit. Here, we present an integrated high-accuracy 3D imaging system combining binocular meta-lens with an optical clue fusion network. Our innovation lies in the synergistic fusion of physics-derived absolute stereo depth measurements and machine learning-estimated relative depth through adaptive confidence mapping - the latter effectively addressing the inherent limitations of absolute depth estimation in scenarios with insufficient matching features. This hybrid approach achieves unprecedented precision of depth estimation (error &lt;1%) while maintaining robust performance across feature-deficient surfaces. The methodology significantly expands viable detection areas and enhances measurement reliability, accelerating practical implementations of metalens-enabled 3D imaging.</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

High-precision three-dimensional imaging based on binocular meta-lens and optical clue fusion

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月1日

查看全文
  • Yuzhou Song,
  • Yifei Zhang,
  • Xiaoyuan Liu,
  • Takuo Tanaka,
  • Mu Ku Chen,
  • Zihan Geng

摘要

Three-dimensional (3D) imaging plays a crucial role in autonomous driving, medical diagnostics, and industrial inspection by providing comprehensive spatial information. Metalens-based 3D imaging is highly valued for imaging applications thanks to its compactness, with enhanced precision remaining a key research pursuit. Here, we present an integrated high-accuracy 3D imaging system combining binocular meta-lens with an optical clue fusion network. Our innovation lies in the synergistic fusion of physics-derived absolute stereo depth measurements and machine learning-estimated relative depth through adaptive confidence mapping - the latter effectively addressing the inherent limitations of absolute depth estimation in scenarios with insufficient matching features. This hybrid approach achieves unprecedented precision of depth estimation (error <1%) while maintaining robust performance across feature-deficient surfaces. The methodology significantly expands viable detection areas and enhances measurement reliability, accelerating practical implementations of metalens-enabled 3D imaging.