基于机器视觉的地下装配式车站监测与安全评估研究

洪成雨, 饶伟, 吴成刚, 戴继, 贾科

doi:10.3973/j.issn.2096-4498.2023.S2.024

隧道建设 >

2023 , Vol. 43 >Issue S2: 215 - 222

DOI: https://doi.org/10.3973/j.issn.2096-4498.2023.S2.024

研究与探索

基于机器视觉的地下装配式车站监测与安全评估研究

展开

（1. 深圳大学土木与交通工程学院，广东深圳 518060； 2. 深圳市地铁地下车站绿色高效智能建造重点实验室，广东深圳 518060； 3. 滨海城市韧性基础设施教育部重点实验室，广东深圳 518060； 4. 北京城建设计发展股份有限公司，北京 100037； 5. 深圳市地铁集团有限公司，广东深圳 518038）

洪成雨(1982—)，男，吉林黑龙江人，2011 年博士毕业于香港理工大学，岩土工程专业，副教授，现从事岩土监测、智能传感器件研发工作。Email： cyhong@szu.edu.cn。*通信作者：饶伟， Email： raowei2021@email.szu.edu.cn。

网络出版日期: 2024-03-27

收起

Machine VisionBased Monitoring and Safety Assessment of Prefabricated Underground Stations

Expand

(1. College of Civil and Transportation Engineering, Shenzhen University, Shenzhen 518060, Guangdong, China; 2. Key Laboratory of Green, Efficient and Intelligent Construction of Shenzhen Metro Underground Station, Shenzhen 518060, Guangdong, China; 3. Key Laboratory of Coastal Urban Resilience Infrastructure, Ministry of Education, Shenzhen 518060, Guangdong, China; 4. Beijing Urban Construction Design and Development Group Co., Ltd., Beijing 100037, China; 5. Shenzhen Metro Group Co., Ltd., Shenzhen, Guangdong 518038, China)

Online published: 2024-03-27

Fold

摘要

为验证机器视觉监测系统在地下工程中的适用性，探明地下装配式车站的沉降变化规律，在现场进行标定试验，并在地下装配式车站的覆土施工和施工后阶段进行沉降监测。将现场监测数据与有限元计算值进行对比，研究建立一套实时的多尺度安全等级评价方法。结果表明： 1）相机与标靶距离小于60 m时，平均测量误差率低于3%，对短距离的结构位移监测具有较高精度。2）实测数据揭示覆土施工作业、出入口开挖作业以及降雨等因素导致车站沉降变化，覆土期间装配式地下车站沉降在-7.0 mm以内，车站结构总体上较为安全。3）机器视觉监测系统实测数据与有限元计算值揭示的车站沉降变化模式一致，机器视觉监测系统能够准确地捕捉和测量地下车站施工过程中的变形情况，在地下工程中具有良好的适用性。

关键词： 地下装配式车站; 机器视觉; 沉降监测; 安全评估

本文引用格式

洪成雨, 饶伟, 吴成刚, 戴继, 贾科 . 基于机器视觉的地下装配式车站监测与安全评估研究[J]. 隧道建设, 2023 , 43(S2) : 215 -222 . DOI: 10.3973/j.issn.2096-4498.2023.S2.024

Abstract

To explore the applicability of the machine vision-based monitoring system in underground engineering and the settlement variation patterns of underground prefabricated stations, field calibration experiments are conducted and the settlement during the construction and postconstruction phases of the underground prefabricated stations is monitored. The field monitoring data is then compared with the finite element calculation results to establish a real-time multi-scale safety level evaluation method. The results show that: (1) When the distance between the camera and the target is less than 60 m, the average measurement error rate is within 3%. This indicates that the system has a high accuracy in monitoring the structural displacements over short distances. (2) The measured data reveal the settlement variations in the stations due to factors such as backfill construction, entrance excavation, and rainfall. (3) During the backfilling phase, the settlement of the prefabricated underground stations is within －7.0 mm, indicating an overall safe condition of the station structure. The measured data obtained from the machine visionbased monitoring system coincide with the settlement variation patterns revealed by the finite element calculation. This demonstrates that the machine visionbased monitoring system can accurately capture and measure the deformations during the construction of underground stations, and it has a good applicability in underground engineering.

Key words： prefabricated underground station; machine vision; settlement monitoring; safety assessment

Options

文章导航

〈

〉

模态框（Modal）标题

摘要

本文引用格式

Abstract