测绘通报 ›› 2022, Vol. 0 ›› Issue (9): 29-33.doi: 10.13474/j.cnki.11-2246.2022.0259

• 交通建设工程测绘技术应用研究 • 上一篇    下一篇

面向地铁隧道表面渗漏水的快速检测技术

田有良1, 樊廷立2, 唐超3   

  1. 1. 南京地铁运营有限责任公司, 江苏 南京 211135;
    2. 武汉大学遥感信息工程学院, 湖北 武汉 430079;
    3. 北京城建勘测设计研究院有限责任公司, 北京 100101
  • 收稿日期:2022-07-11 修回日期:2022-08-02 发布日期:2022-09-30
  • 通讯作者: 樊廷立。E-mail:fantingli1990@163.com
  • 作者简介:田有良(1982—),男,硕士,主要从事地铁结构变形与保护工作。E-mail:93544860@qq.com

Rapid detection technology for surface water leakage in subway tunnel

TIAN Youliang1, FAN Tingli2, TANG Chao3   

  1. 1. Nanjing Metro Operation Co., Ltd., Nanjing 211135, China;
    2. School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China;
    3. Beijing Urban Construction Exploration & Surveying Design Research Institute Co., Ltd., Beijing 100101, China
  • Received:2022-07-11 Revised:2022-08-02 Published:2022-09-30

摘要: 随着地铁运营时间的不断增加、地下水位的上涨,地铁隧道渗漏水情况日益严重,已严重影响地铁隧道的安全运行。传统的检测方法为人工现场巡查,效率低、准确率差,高自动化、高精度、高稳定性的漏水检测方法是改进传统检测方法的关键。因此,本文提出了一种利用移动激光扫描隧道进行渗漏水检测的深度学习方法。该方法由以下部分组成:①利用获取的隧道衬砌点云建立渗漏水数据集;②通过基于掩码区域卷积神经网络进行自动渗漏检测。在南京地铁2号线奥体东—兴隆大街测试结果表明,本文方法实现了隧道衬砌漏水在二维平面的自动化检测和评价,为检测人员提供了直观的漏水信息展示。

关键词: 地铁隧道, 三维激光, 深度学习, 灰度图转换, 图像二值化

Abstract: With the continuous increase of subway operation time and the rise of underground water level, the leakage of subway tunnel is becoming more and more serious, which has seriously affected the safe operation of subway tunnel. The traditional detection method is manual field inspection, which has low efficiency and poor accuracy. The water leakage detection method with high automation, high accuracy and high stability is the key to improve the manual detection method. Therefore, this paper proposes a deep learning method for water leakage detection using mobile laser scanning tunnels. The method consists of the following parts.①Water leakage data set is established by using the obtained tunnel lining point cloud.②Automatic leak detection is carried out by convolutional neural network based on mask region.The test results of Aoti East-Xinglong street of Nanjing metro line 2 show that the proposed method can realize automatic detection and evaluation of tunnel lining water leakage in two-dimensional plane, and provide intuitive display of water leakage information for inspectors.

Key words: subway tunnel, 3D laser, deep learning, grayscale transformation, image binarization

中图分类号: