桥梁全息性态特征提取方法试验研究

Experimental Study on the Holographic Feature Extraction Method of Bridge

为了克服传统接触式位移测量只能获取结构有限离散点的位移数据,难以准确、连续地表征整个结构几何变形的缺点。在统计学归一化互相关分析的基础上提出一种无需在结构上布置任何靶点就能快速从结构振动视频中获取结构表面密集特征点位移响应的灰度值匹配算法。以实验室内超大跨自锚式悬索桥模型为研究对象,依次开展了小车行驶激励和单点激励两类试验,通过对同一视域内单目标和多目标追踪验证了本文算法的准确性、可行性和高效性,并提出匹配模板概念将基础算法做了进一步优化,进而缩短计算时间。试验结果表明:本文方法与传统接触式位移传感器获取结构动态位移响应的误差在5%以内,精度满足工程实践要求。本文算法可以在无靶点的情况下获取结构表面任意特征点的动态位移响应,且具有较高的精度、效率和可靠性,相对于有靶点非接触摄影测量和常规接触式位移测量,研究结果可操作性和普适性更强,获取的数据更加全面。

The traditional contact displacement measurement can only obtain the displacement data of the finite discrete points of the structure,and it is difficult to characterize the geometric deformation of the whole structure accurately and continuously.To overcome the shortcomings,based on the statistical normalized cross correlation analysis,a gray value matching algorithm was proposed.It could obtain the displacement response of the dense feature points on the structure surface from the vibration video of the structure without arranging any targets on the structure quickly.Taking the model of super-span self-anchored suspension bridge in the laboratory as the research object,two kinds of tests were carried out successively.One was driving excitation,and the other was single point excitation.The accuracy,feasibility and efficiency of the algorithm were verified by tracking single target and multi-target in the same field of view,and the concept of matching template was proposed to further optimize the basic algorithm,so as to shorten the calculation time.The test results show that the error between the proposed method and the traditional contact displacement sensor is less than 5%,and the accuracy meets the engineering practice requirements.The algorithm can obtain the dynamic displacement response of any feature point on the surface of the structure without target.And it has high precision,efficiency and reliability.Compared with target non-contact photogrammetry and conventional contact displacement measurement,the proposed method is of much operability and universality,and the acquired data is more comprehensive.