摘要
针对桩基检测环境复杂,存在复杂噪声的问题.提出一种基于峭度准则和信号相关性分析的经验模态分解(EMD)去噪改进算法,通过峭度准则判断分量包含特征信息的多少,通过信号相关性分析判断分量是否为噪声分量,利用改进EMD时空去噪方法对信号进行第一次去噪,在此基础上,利用EMD小波阈值法对信号进行二次去噪,两次去噪构成本文完整的基于EMD的信号去噪改进算法.实验仿真显示该去噪方法对于桩基微弱信号在保留信号特征信息的前提下取得了显著的去噪效果.信号的去噪效果将直接影响到桩基缺陷信号的识别检测,具有重要的应用价值和研究意义.
For pile foundation detection complex environment,there are complex noise problems.This paper proposed an improved algorithm based on kurtosis criterion and information correlation analysis for EMD denoising.Based on the initial denoising,the EMD wavelet threshold method was used to further denoise the signal.The denoising twice constitutes the complete signal denoising process.The experimental simulation showed that the denoising method had achieved a significant denoising effect on the premise of retaining the weak signal of the pile foundation on the premise of retaining the signal feature information.
作者
李衡
康维新
LI Heng;KANG Wei-xin(School of Information and Communication Engineering,Harbin Engineering University,Harbin 150001,China)
出处
《哈尔滨商业大学学报（自然科学版）》
CAS
2019年第1期81-85,共5页
Journal of Harbin University of Commerce：Natural Sciences Edition
基金
国家自然科学基金(61371174)
关键词
经验模态分解
峭度准则
自相关分析
桩基
无损检测
去噪
EMD小波阈值法
EMD
kurtosis index
correlation analysis
pile foundation
non-destructive testing
de-noising
EMD wavelet threshold method
