摘要
边坡工程监测数据往往包含各种误差(噪声),表现出明显的非真实波动和突变,边坡的实际变形特征被噪声淹没而无法识别。小波分析能利用边坡真实变形信号与噪声的时频特性不同,借助小波变换的多分辨率分析,有效地对不同频率成分进行分离,并通过作用阈值,最终达到降噪的目的并提取出边坡真实的变形特征。工程中获得的监测数据有时是非等间隔的,不能直接进行小波降噪,可以利用三次厄密插值使其转变为等间隔数据序列,分析结果表明这种处理方法是可行的。
Monitoring data in slope engineering often contain a variety of errors（noise）,and shows obvious fluctuation and mutation,so that the deformation features of the slope have been submerged in noise and can not be identified.Wavelet analysis can take advantage of the time-frequency difference between the real slope deformation signal and the noise,with the help of wavelet transform multi-resolution analysis,to effectively separate different frequency components and finally de-noise and find the real characteristics of slope deformation after a combined work on threshold.Monitoring data in the project are sometimes unequally spaced which can not be directly used for wavelet de-noising.However they can be transferred to equally spaced data sequence by cubic Hermite interpolation.The results proves that the feasibility of this approach.
出处
《公路工程》
2011年第1期16-19,共4页
Highway Engineering
基金
国家自然科学基金资助项目(40972187)
关键词
小波降噪
非等间隔数据序列
厄密插值
边坡监测
wavelet de-noising
unequally spaced data sequence
Hermit interpolation
slope monitoring
