基于改进LSD的斜拉桥索力非接触鲁棒识别  

Robust Non-contact Recognition of Cable Forces in Cable-stayed Bridges Based on Improved LSD

作　　者：朱前坤[1] 崔亚歌 王宪玉 杜永峰[1] ZHU Qiankun;CUI Yage;WANG Xianyu;DU Yongfeng(Institute of Earthquake Protection and Disaster Mitigation,Lanzhou University of Technology,Lanzhou 730050,China;Gansu Province Transportation Planning Survey and Design Institute Co.,Ltd.,Lanzhou 730030,China)

机构地区：[1]兰州理工大学防震减灾研究所,甘肃兰州730050 [2]甘肃省交通规划勘察设计院股份有限公司,甘肃兰州730030

出　　处：《湖南大学学报（自然科学版）》2024年第11期158-166,共9页Journal of Hunan University:Natural Sciences

基　　金：国家自然科学基金资助项目(52168041);甘肃省重点研发计划资助项目(22YF11GA301)。

摘　　要：拉索是斜拉桥的关键承载构件,其受力状态是评估桥梁安全状态的重要指标,准确测量拉索索力是保障桥梁结构安全的重要环节.基于此,建立一套可实现远距离、无靶标、高精度的斜拉索索力识别系统.该系统利用改进的直线检测(Line Segment Detector,LSD)算法对拉索进行追踪,从而识别拉索的振动信息;再通过模态分解(VMD)离散出斜拉索的有效振动信号,得到斜拉索的频率信息;最后利用振动频率法,从斜拉索的频率信息来估算出斜拉索的索力.在实验室条件下对长细比较大、像素占比较少的钢尺模型以及在大跨度斜拉桥上对斜拉索索力进行识别对比,确定了改进的LSD算法以及该系统对索力识别的精确性及鲁棒性.在钢尺模型试验中,改进的LSD算法识别到的钢尺振动信息与激光位移计对比误差仅为0.63%;在大跨度斜拉桥上的试验中,该系统识别到斜拉索的索力与接触式传感器对比误差小于3.0%.以上两个试验结果表明,本文提出的系统能够在复杂的实际工程现场环境中对斜拉索索力进行高精度识别.The cable is a key load-bearing component of cable-stayed bridges,and its stress state is an important indicator for assessing the safety of bridge structures.Accurate measurement of cable forces is crucial to ensure the safety of the bridge.Based on this,this paper establishes a set of cable force identification system for cable-stayed bridges that can achieve long-distance,non-target,and high-precision measurements.The system utilizes an improved line segment detector(LSD)to track the cables and identify their vibration information.The valid vibration signals of the cables are then discretely extracted using the variational mode decomposition(VMD)method to obtain their frequency information.Finally,the cable forces are estimated using the vibration frequency method based on the cable’s frequency information.The accuracy and robustness of the improved LSD and the cable force identification system are confirmed through experiments conducted in the laboratory on a model with a large slenderness ratio and low pixel ratio,as well as cable force identification on a large-span cable-stayed bridge.In the laboratory experiment using a steel ruler model,the improved LSD identified the vibration information of the ruler with an error of only 0.63%compared with the laser displacement sensor.In the experiment on a large-span cable-stayed bridge,the error in the identified cable forces by the system compared with the contact sensors was less than 3.0%.Both experiments demonstrate that the proposed system can accurately identify cable forces in complex engineering field environments.

关 键 词：桥梁工程 索力测量系统 计算机视觉 直线检测 无靶标识别 

分 类 号：U448.27[建筑科学—桥梁与隧道工程]