经验模式分解的尺度阈值滤波在变形分析中的应用  

EMD methods' application in deformation analysis

作　　者：余春平[1] 李广云[1] 卫建东[1] 张皓[1] 

机构地区：[1]解放军信息工程大学测绘学院,郑州450052

出　　处：《测绘科学》2008年第6期189-191,共3页Science of Surveying and Mapping

摘　　要：本文将信号处理领域的经验模式分解算法应用于变形信息的提取中,通过引入阈值函数,建立了基于经验模式分解的尺度阈值滤波模型,采用优化模型确定了经验模式分解的次数。分别通过模拟试验和实测数据与小波阈值法和多项式拟合法进行了比对,分析表明:在低噪声情况下,三种方法都有一定的滤波效果;在高噪声情况下,经验模式分解的尺度阈值滤波法具有与小波阈值去噪法相等的精度,而且瞬时强噪声识别能力更好,优于多项式拟合法。The deformation monitoring is aim to ensure the safe of the structure, but in fact, the real deformation information is usually hidden in a mount of noises. So it＇ s of great significance to recognize the useful deformation information with the right data processing method for the promotion of deformation analysis accuracy. Through bringing the Empirical Mode Deformation and threshold function to the field of deformation analysis, this paper proposes a scale threshold filtering method and fixes on the decomposing degree. With the simulated and survey data, the author do a series of comparisons through EMD scale filter, wavelet transform and polynomial fitting, draw some conclusions： At lower noise level, each method＇ result is nice; at the higher noise level, EMD Scale Filter and wavelet transform get better result than polynomial fitting. At last, some suggestions are given according to the different feature of methods.

关 键 词：经验模式分解 EMD尺度滤波 阈值化 变形分析 

分 类 号：P258[天文地球—测绘科学与技术]