中国建设信息化2024，Issue(23) ：58-61.

基于深度学习与数字孪生技术的建筑钢结构检测方法

Construction Steel Structure Inspection Method Based on Deep Learning and Digital Twin Technology

孙晓强 刘皓宇 沙奕 张天辉
中国建设信息化2024，Issue(23) ：58-61.
✕
  • 维普
  • 万方数据

基于深度学习与数字孪生技术的建筑钢结构检测方法

Construction Steel Structure Inspection Method Based on Deep Learning and Digital Twin Technology

孙晓强 1刘皓宇 1沙奕 1张天辉1
扫码查看

作者信息

  • 1. 北京城建集团有限责任公司
  • 折叠

摘要

针对传统建筑钢结构缺陷检测方法效率较低的问题,研究通过特征金字塔网络和逐元素加法对快速区域卷积神经网络进行改进,并将其与数字孪生技术结合,提出了一种基于深度学习与数字孪生技术的建筑钢结构检测方法.结果显示,该方法在不同尺寸下的检测精度分别为0.78、0.81、0.82,证明了其可靠性较高.表明设计的方法能够准确地识别钢构件缺陷,研究结果可应用于建筑物结构健康监控领域,为钢结构风险预警和维护提供有力的技术支持.

Abstract

To address the issue of low efficiency in traditional methods for detecting defects in building steel structures,this research improves the Fast Region Convolutional Neural Network(Fast R-CNN)by using a Feature Pyramid Network(FPN)and element-wise addition.By integrating this with digital twin technology,a detection method for building steel structures based on deep learning and digital twin technology is proposed.The results show that the detection accuracy for different sizes is 0.78,0.81,and 0.82,respectively,proving its high reliability.This indicates that the designed method can accurately identify defects in steel components.The research findings can be applied in the field of structural health monitoring of buildings,providing strong technical support for risk warning and maintenance of steel structuresa.

关键词

建筑钢结构/缺陷检测/快速区域卷积神经网络/数字孪生技术

Key words

Building Steel Structure/Defect Detection/Fast Region Convolutional Neural Network/Digital Twin Technology

引用本文复制引用

出版年

2024
中国建设信息化

中国建设信息化

ISSN：
段落导航相关论文