基于视觉与振动数据融合的结构层间位移响应识别  

Interstory drift response identification of structures based on vision and vibration data fusion

作　　者：单伽锃 张宏泽 白子轩 周德源[1] SHAN Jiazeng;ZHANG Hongze;BAI Zixuan;ZHOU Deyuan(College of Civil Engineering,Tongji University,Shanghai 200092,China;Shanghai Engineering Research Center for Resilient Cities and Intelligent Disaster Mitigation,Shanghai 200092,China)

机构地区：[1]同济大学土木工程学院,上海200092 [2]上海韧性城市与智能防灾工程技术研究中心,上海200092

出　　处：《东南大学学报(自然科学版)》2025年第1期41-50,共10页Journal of Southeast University:Natural Science Edition

基　　金：国家自然科学基金资助项目(52278312);中央高校基本科研业务费专项资金资助项目。

摘　　要：利用基于计算机视觉的接触式位移识别技术分析结构振动视频来捕捉结构层间位移响应,可以解决极端荷载作用下结构层间位移观测困难的问题。结合结构层间变形机理与多特征检测算法,提出了结构弯剪耦合变形识别方法,并在理论上验证了该方法在结构发生多种变形情况下的适用性。该方法利用倾角仪数据与视觉识别结果计算得到结构的层间位移、有害层间位移、无害层间位移及部分位移在总位移中的占比,并在缩尺模型振动台试验中验证了该方法的准确性与有效性。研究结果表明,该方法可用于建筑结构在地震过程中的关键变形监测,即通过结构不同层间变形成分的比例可以准确判断结构各层实际受力状态。该方法通过多数据计算显著提高了层间位移的识别精度,为建筑层间受力状态评估提供了重要依据。With a contact-based displacement recognition technology based on computer vision,structural vi-bration videos can be analyzed to capture interstory displacement responses in structures,addressing the chal-lenges of observing interstory displacements under extreme loading conditions.By integrating the mechanisms of structural interstory drift with multi-feature detection algorithms,a method for recognizing bending-shear coupled deformations is proposed.The theoretical validation demonstrates the method’s applicability under various deformation scenarios.This technique employs data from inclinometers and visual recognition results to compute interstory drifts,harmful drifts and harmless drifts,thus providing proportions of different drifts within the total drift.Validation through scaled model shaking table experiments confirms the accuracy and ef-fectiveness of this method.The findings indicate that this approach can be utilized for critical deformation monitoring in buildings during seismic events,enabling an accurate assessment of the actual stress states of dif-ferent stories based on the proportions of their deformation components.This approach significantly enhances the precision of interlayer displacement recognition by multi-data calculation,providing essential evidence for evaluating the stress states within building layers.

关 键 词：结构健康监测 变形识别 层间位移 计算机视觉 

分 类 号：TU196.4[建筑科学—建筑理论]