Scholarlink

<p>With the rapid development of transportation infrastructure, effective road management and maintenance have become increasingly important. Pavement detection is critical for improving road quality, ensuring traffic safety, and promoting sustainable development. However, traditional 2D detection methods, while capable of providing pavement information, are limited in dimensionality and fail to fully capture the 3D characteristics of pavement surfaces. On highways, achieving an acquisition speed of 100&#xa0;km/h presents challenges for single-camera systems, which struggle to cover a 4-m-wide lane while maintaining 1&#xa0;mm measurement precision, particularly in complex environments. To address these limitations, this study introduces three key innovations: Hardware optimization: (1). A dual-camera oblique layout is implemented to expand the detection range, achieving measurement accuracy better than 1&#xa0;mm under high-speed conditions. (2). Zigzag calibration algorithm: a specially designed zigzag 3D calibration block extracts feature points in complex environments, enabling precise dual-camera calibration and improving system stability. (3). Multi-camera 3D reconstruction: calibration files facilitate image stitching across multiple cameras, ensuring geometric consistency and providing distortion-free panoramic lane views through advanced rendering techniques. The proposed system ensures precision better than 1&#xa0;mm in the X, Y, and Z directions at high speeds, delivering complete and distortion-free lane coverage. This work provides critical technical support for wide-field 3D structured light scanning systems, offering insights for intelligent, high-speed, and high-precision pavement detection. It also contributes to the digital transformation of road maintenance and management.</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-Speed Sub-millimeter 3D Pavement Data Acquisition: Calibration, Reconstruction, and Height Correction

  • 期刊论文

  • 发表日期：

    2025年6月2日

查看全文
  • Ao Dai,
  • Sijiang Heng,
  • Lin Li,
  • Wenting Luo,
  • Wanheng Li,
  • Haizhu Lu,
  • Juncheng Zeng,
  • Yan Zong,
  • Chao Zhang

摘要

With the rapid development of transportation infrastructure, effective road management and maintenance have become increasingly important. Pavement detection is critical for improving road quality, ensuring traffic safety, and promoting sustainable development. However, traditional 2D detection methods, while capable of providing pavement information, are limited in dimensionality and fail to fully capture the 3D characteristics of pavement surfaces. On highways, achieving an acquisition speed of 100 km/h presents challenges for single-camera systems, which struggle to cover a 4-m-wide lane while maintaining 1 mm measurement precision, particularly in complex environments. To address these limitations, this study introduces three key innovations: Hardware optimization: (1). A dual-camera oblique layout is implemented to expand the detection range, achieving measurement accuracy better than 1 mm under high-speed conditions. (2). Zigzag calibration algorithm: a specially designed zigzag 3D calibration block extracts feature points in complex environments, enabling precise dual-camera calibration and improving system stability. (3). Multi-camera 3D reconstruction: calibration files facilitate image stitching across multiple cameras, ensuring geometric consistency and providing distortion-free panoramic lane views through advanced rendering techniques. The proposed system ensures precision better than 1 mm in the X, Y, and Z directions at high speeds, delivering complete and distortion-free lane coverage. This work provides critical technical support for wide-field 3D structured light scanning systems, offering insights for intelligent, high-speed, and high-precision pavement detection. It also contributes to the digital transformation of road maintenance and management.