基于ITD和改进小波阈值的超声检测信号去噪  

Improved Threshold De-noising Method of Ultrasonic Test Signal Based on ITD and Wavelet Transform

作　　者：李大中[1] 赵杰[1] 刘建屏[2] 蔡文河[2] 马延会[2] 

机构地区：[1]华北电力大学,河北保定071003 [2]国网冀北电力有限公司电力科学研究院(华北电力科学研究院有限责任公司),北京100045

出　　处：《华北电力技术》2016年第9期8-13,共6页North China Electric Power

摘　　要：超声检测信号是一种非线性、不平稳信号,且由于材料结构的复杂性,超声回波信号往往存在很多干扰噪声。针对钢制结构中平底孔的超声检测信号传统小波去噪方法的不足,提出固有时间尺度分解(ITD)和小波阈值去噪相结合的方法对信号进行去噪处理。通过将带噪声的原始检测信号进行ITD分解,获得一系列固有旋转分量(PRC)和一个单调趋势项,去除高频分量后重构部分主要的PRC分量,然后采用改进阈值的小波软阈值方法对其进行去噪,该方法融合了小波去噪和ITD方法的优势,信噪比、均方根误差和相关性等参数都比较满意,去噪效果明显。The ultrasonic detect signal is a nonlinear and unstable signal,and because of the complexity of the structure of the material,the ultrasonic echo signal often has a lot of interference noise. The method combined Intrinsic time-scale decomposition（ ITD） and wavelet threshold de-noising was proposed for signal de-noising by analyzing the disadvantages of traditional wavelet threshold de-noising methods for flat bottom hole ultrasonic testing signals in the steel structure. The original signal with noise was decomposed into several proper rotation components and a trend component by the ITD method. After removing the high frequency component,part of PRC component was refactored.Wavelet soft-threshold method of improved threshold was applied on the de-noising of it. The method combines the advantage of wavelet de-noising and ITD. The parameters of SNR,RMSE and correlation is quite satisfied,and denoising effect is obvious.

关 键 词：超声检测 去噪 固有时间尺度分解 小波 阈值 

分 类 号：TN911.72[电子电信—通信与信息系统]