中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊
重庆交通大学学报（自然科学版）

重庆交通大学学报（自然科学版） ›› 2020, Vol. 39 ›› Issue (08): 14-19.DOI: 10.3969/j.issn.1674-0696.2020.08.03

• 交通+大数据人工智能 • 上一篇    下一篇

基于地面三维激光扫描的高速公路沉陷量自动提取与分析研究

赵立都1，陈波2，刘国强2, 吴逸飞2，周银1 , 杨宇鹏3，郭彤4，韩达光1   

  1. (1. 重庆交通大学 土木工程学院，重庆 400074； 2. 重庆建工集团 四川遂资高速公路有限公司，四川 遂宁629000； 3.中铁十六局 第五工程公司，宁夏 银川 751400; 4. 东南大学 土木工程学院，江苏 南京 210096)
  • 收稿日期:2018-10-26 修回日期:2019-05-05 出版日期:2020-08-18 发布日期:2020-08-25
  • 作者简介:赵立都（1991—），男，甘肃庆阳人，博士研究生，主要从事桥梁健康监测与智慧交通基础设施应用技术方面的研究。E-mail:compassgps@163.com 通信作者：韩达光（1982—），男，辽宁沈阳人，高级工程师，博士，主要从事BIM与土木工程融合方面的研究。E-mail:daguang.han@cqjtu.edu.cn
  • 基金资助:
    重庆建工集团遂宁高速公路公司BIM+GIS关键技术研发课题

Automatic Extraction and Analysis of Expressway Subsidence Based on Ground 3D Laser Scanning

ZHAO Lidu1, CHEN Bo2, LIU Guoqiang2, WU Yifei2, ZHOU Yin1, YANG Yupeng3, GUO Tong4, HAN Daguang1   

  1. (1. College of Civil Engineering, Chongqing Jiaotong University, Chongqing 400074, China; 2. Sichuan Suizi Expressway Co. Ltd., Chongqing Construction Engineering Group, Suining 629000, Sichuang, China; 3. Fifth Engineering Company, China Railway Sixteenth Bureau, Yinchuan 751400, Ningxia, China; 4. School of Civil Engineering, Southeast University, Nanjing 210096, Jiangsu, China)
  • Received:2018-10-26 Revised:2019-05-05 Online:2020-08-18 Published:2020-08-25

PDF

摘要： 高速公路运维过程中，路面沉陷是影响行车安全的关键，传统多采用水准测量等技术方法，存在工作量大、周期长、易受外界条件影响等缺点。提出一种通过地面三维激光扫描点云自动判别路面沉陷的算法和分析方法，该方法先采用基于几何特征与迭代最邻近点算法相结合的方法进行点云配准，配准数据经过点云分类识别后，基于滑动窗口法提取道路特征线，建立高速公路的路面模型，并通过分析该模型得到沉陷区域、位置和沉陷深度等指标。为验证算法的可靠性与准确性，以一段运营维护阶段的高速公路为研究对象进行分析实验。通过算法处理与实测路面模型分析，发现距离高速公路实验段起点45 m处存在深度达40 cm的沉陷区。分析结果表明，采用提出的对激光点云数据处理算法和路面模型分析方法，可以高效、准确分析高速公路路面的沉陷情况。该方法的成功应用，对三维激光扫描技术的在交通领域的深度应用和有效推广有重要意义。

关键词: 道路工程, 地面三维激光扫描, 点云, 道路沉陷监测, 道路特征线提取

Abstract: The road surface settlement is the key to influence the safety of driving during the operation and maintenance of expressways. Conventional inspections are mostly based on leveling, which have the disadvantages of heavy workload, long cycle and easy to be affected by external conditions. An algorithm and analysis method for automatically identifying pavement subsidence by ground 3D laser scanning point cloud was proposed. In the proposed method, the point cloud was registered firstly based on the combination of geometric features and iterative closest point algorithm. After the registration data was classified and identified by point cloud, the road feature line was extracted based on the sliding window method and the pavement model of the expressway was established. By analyzing the proposed model, the indexes of subsidence area, location and depth were obtained. Finally, a highway at the operation and maintenance stage was adopted to verify the feasibility and accuracy of the proposed algorithm. Through the algorithm processing and the actual pavement model analysis, it is found that there is a settlement area with a depth of 40 cm at 45 m away from the starting point of the experimental section of the expressway. The analysis results show that the proposed algorithm of laser point cloud data processing and pavement model analysis can efficiently and accurately analyze the settlement of expressway pavement. The successful application of the proposed method is of great significance for the deep application and effective promotion of 3D laser scanning technology in the traffic field.

Key words: highway engineering, ground 3D laser scanning, point clouds, road settlement monitoring, road feature line extraction

中图分类号: