摘要
针对微震信号能量弱、信噪比低的问题,本文提出了一种基于多重同步压缩变换(MSST)与小波阈值“强强联合”的微震信号去噪方法,实现微震信号噪声压制.该方法主要利用小波阈值对多重同步压缩变换系数进行去噪,进而反变换实现信号重构,在仿真信号和实际微震单道记录中均得到验证.仿真结果表明,相较于短时傅里叶变换(STFT)、连续小波变换(CWT)、同步压缩变换(SST),MSST具有更好的时频系数浓缩能力,瑞利熵最小;本文方法信噪比提高3dB,均方误差最小(MSE=0.12),峰值信噪比PSNR最大(PSNR=18.29),平滑度指标SM较小,且计算时间短,仅为50ms,远远小于高阶SST,具有明显的优势.在此基础上,将此方法应用于实际微震单道记录,测试结果表明该方法具有较好的噪声压制能力,具有实用价值.
In view of the problems of weak energy and low SNR of microseismic signals,a microseismic signal denoising method based on the“strong combination”of multiple synchronous compression transform(MSST)and wavelet threshold was proposed to suppress the noise of the microseismic signal.This method mainly used wavelet threshold to denoise multiple synchronous compression transform coefficients,and then inversed transforms to realize signal reconstruction,which had been verified in both simulated signal and actual microseismic single-track recording.The simulation results showed that,compared with the STFT,CWT,and SST,MSST had better time-frequency coefficient concentration ability and the smallest Rayleigh entropy.The proposed method had a 3 dB improvement in SNR,the smallest MSE(which was 0.12),the largest PSNR(which was 18.29),a smaller SM index,and a short computation time of 50 ms,which was much less than that of the high-order SST,indicating that the method has obvious advantages.On this basis,this method was applied to the actual microseismic single-track recording.The test results showed that this method has good noise suppression ability and has practical value.
作者
李亦佳
王静
王正方
隋青美
LI Yijia;WANG Jing;WANG Zhengfang;SUI Qingmei(School of Control Science and Engineering,Shandong University,Jinan 250061,China;Jiangsu Hengkang Electromechanical Co.,Ltd.Yancheng 224014,China)
出处
《应用基础与工程科学学报》
EI
CSCD
北大核心
2022年第2期486-500,共15页
Journal of Basic Science and Engineering
基金
山东省自然科学基金面上项目(ZR2018MEE052)
国家自然科学基金项目(41877230)
山东省重点研发计划项目(2019GGX01027)
国家自然科学基金联合项目(U1806226)。
关键词
噪声压制
多重同步压缩变换
时频分布
瑞利熵
信号重构
noise suppression
multisynchrosqueezing transform
time-frequency distribution
Renyi entropy
signal reconstruction
