【作者】 赵强；

【导师】 钟佑明；

【作者基本信息】 重庆交通大学 ， 计算机应用技术， 2011， 硕士

【摘要】 裂缝是桥梁健康状态的重要征兆之一,也是桥梁安全状态恶化的重要根源之一,对其进行实时监测具有十分重要的意义。随着视频扫描技术、计算机技术、无线传输技术和信号分析技术的迅猛发展,裂缝监测呈现出新的面貌。本文主要对导师及其合作者提出的一款桥梁裂缝远程监测系统——BCMSBVF中的一些关键技术进行了研究和解决,实现了该系统的几项重要功能,并对它们的有效性进行了验证。本文系统中采用了一种新的信号分解方法——经验模态分解法(Empirical Mode Decomposition,EMD)来进行裂缝图片处理。EMD是Hilbert-Huang变换(Hilbert-Huang Transform,HHT)信号处理方法的核心技术,它直接针对数据,可以自适应地将一维信号分解成有限个固有模态函数(Intrinsic Mode Function,IMF)。由于EMD分析方法是基于数据的局部特征尺度的,而不是其它既定函数(如傅立叶变换中的正余弦函数),因而EMD方法既可以适用于平稳信号,又能运用于非平稳信号。EMD有几项关键技术——包络线拟合技术、终止准则技术、端点效应抑制技术等,这些技术的好坏决定了EMD结果的可靠性和有效性。针对EMD的端点效应抑制问题,本文成功实现了一种基于本征波匹配的抑制方法——本征波匹配预测法,并用巧妙构造的仿真信号和实测的实际信号算例验证了本征波匹配预测法的有效性和先进性。桥梁裂缝监测需要对图像信号进行实时在线处理,但目前还没有实时EMD方法。针对这一问题,本文使用信号分段、均值继承、重叠处理等几项特别技术提出了实时EMD处理算法——重叠分段-均值继承EMD法(简称SS-MIEMD),从而实现了对在线信号的动态、连续、快速、高精度EMD处理。通过若干仿真和实测信号算例验证了重叠分段-均值继承EMD法的有效性和可行性,并将该算法扩展至了二维EMD方法。同时,本文将嵌入式技术、GPRS技术引入到BCMSBVF系统,构建并初步实现了一款基于GPRS和HHT的嵌入式桥梁裂缝远程监测系统,该系统结构简洁、使用灵活,可以实现对桥梁视频图片的远程采集、裂缝识别、信息发布和灾害报警等功能。在实验室环境下进行了实验,验证了该系统的效果。更多还原

【Abstract】 Crack is one of important signs of a bridge health status, and it is also a major source of bridge security status deterioration, so it has great significance to real-time monitor it. With a rapid development of video scanning technology, computer technology, wireless transmission technology and signal analysis technology, a new look is shown in crack monitoring. Some key technologies of a remote monitoring system for bridge crack proposed by author’s tutor and his collaborators are studied in this paper. Several key features of the system are achieved, and their effectiveness is verified.A new signal decomposition method - empirical mode decomposition method (EMD) is used in the system for crack image processing. EMD is the core technology of signal processing method - Hilbert-Huang transform (HHT), it directly against the data and can adaptively decompose one-dimensional signal into finite number of intrinsic mode function (IMF). The EMD method is based on the scale of data’s local characteristic, rather than other established functions (such as cosine function in Fourier transform), and thus the EMD method can be applied both to stationary signals and non-stationary signal.There are several key technologies in EMD - fitting envelope, termination criteria technology, end effect suppression technology etc. The quality of these techniques determines the reliability and validity of EMD. To suppress EMD end effect, a new inhibition method– the intrinsic wave matching and prediction method is successfully implemented in this paper based on intrinsic wave matching. The effectiveness and advancement of the method is illustrated by cleverly constructed simulated signal and actual measured signal.Real-time on-line image signal processing is required in bridge crack monitoring, but there is no real-time EMD method at present. To solve this problem, we present a real-time EMD method - overlapping segments - means inheritance EMD method (referred to as the SS-MIEMD) by using several special techniques such as signal segment, mean inheritance, overlap processing etc., A dynamic, continuous, rapid and high-precision EMD treatment can be achieved by this method. The effectiveness and feasibility of the method is validated by a number of simulated measured signals, the algorithm is also extended to the two-dimensional EMD method.At the same time, the embedded technology, GPRS technology is introduced into BCMSBVF and an embedded remote crack monitoring system of bridge based on GPRS and EMD is constructed and initial realized in the paper. The system has simple structure and is flexible to use. It can achieve the function of remote bridge video image collection, information dissemination and disaster alarm. The experiment under laboratory condition verifies the effectiveness of the system.更多还原