基于混沌特性的滑坡监测序列的小波去噪  

The Sequences of Landslide Monitoring of Wavelet De-noising Based on Chaotic Characteristics

作　　者：郑佳[1] 文鸿雁[1] 袁昌茂[2] 李超[1] 

机构地区：[1]桂林理工大学,广西桂林541004 [2]广东博罗县国土资源局,广东惠州516100

出　　处：《城市勘测》2011年第3期109-112,共4页Urban Geotechnical Investigation & Surveying

基　　金：国家自然科学基金项目(41071294);广西区应用基础研究专项(桂科基0991023);广西研究生教育创新计划资助项目(201005960816M19)

摘　　要：在滑坡监测数据的分析中,噪声滤波是一项重要的工作,特别是滑坡变形系统处于混沌态时,混杂在混沌动力系统中的噪声信号,会掩盖系统的内在动力学特性,影响对系统的准确分析和预报。通过对监测混沌序列和噪声的小波分析,滑坡监测混沌信号的小波变换具有自相似性和噪声随分解尺度变化特征,混沌性质信号的小波变换和原信号在一定的尺度范围内具有标度不变性,具有相同的标度指数;原信号的小波变换受噪声的影响随着尺度的增大而减小,当尺度足够大时,噪声的影响几乎完全消失,因而我们可以通过选择合适的小波基函数和分解层数实现最佳的去噪。根据监测序列的先验信息,选择合适的阈值处理小波分解的系数,这样就能够在小波重构过程达到滤波的效果,仿真实例证明这种方法可有效地去除滑坡监测数据中的噪声。In the landslide monitoring data analysis,the noise filtering is an important work,especially landslide in the state of chaos system,the noise signal mixed in chaotic system,which will cover the inner dynamics characteristic of the system,and affect accurately analysis and prediction of the system.Based on the wavelet analysis of chaotic sequences and noise of landslide monitoring,chaotic signal of the landslide monitoring has self-similarity with wavelet transform and noise characteristics change with scale of decomposition,chaotic signal with wavelet transform have scale invariance in certain scale with the original signal,and have the same standard index;the wavelet transform of original signal effecting by noise reduce with the scale decreases,if the scale is big enough,the noise affect almost completely disappear,thus we can select suitable wavelet function and decomposition layers,eliminating noise to achieve the best.According to the sequence of monitoring information,process the wavelet coefficient with the suitable threshold,which can filter the noise in the wavelet-reconstruction,the simulation results show the method can effectively eliminate the noise of landslide monitoring data.

关 键 词：滑坡监测 混沌 去噪 LYAPUNOV指数 功率谱 

分 类 号：TU196[建筑科学—建筑理论]