随钻地震参考信号稀疏分解去噪  

Denoising with sparse decomposition for the reference signal of seismic while drilling

作　　者：秦显科 王林飞[1] 张博伦[1] 

机构地区：[1]中国海洋大学海洋地球科学学院,山东青岛266100

出　　处：《中国科技论文》2016年第21期2473-2478,共6页China Sciencepaper

基　　金：国家自然科学基金资助项目(41204087);高等学校博士学科点专项科研基金资助项目(20120132120030)

摘　　要：针对随钻地震资料信噪比较低的难题,将稀疏表示理论引入随钻地震参考信号的处理。根据希尔伯特变换求取实际随钻地震参考信号的瞬时属性,使用Morlet子波,构建合适的过完备原子字典,对随钻地震参考信号进行稀疏分解,采用L1范数谱投影梯度算法实现基追踪降噪。利用该方法实现了对实际随钻地震参考信号的稀疏分解降噪,改善了资料的信噪比,恢复了钻头震源信号。To solve the problem of drilling seismic data with low signal to noise ratio, sparse representation theory is introduced into the reference signal processing for seismic while drilling （SWD）. According to the Hilbert transform, the instantaneous at- tributes of the actual seismic reference signal are obtained, the over-complete dictionary of atoms was built using the Morlet wave- let, the reference signal for SWD was sparsely decomposed. Then the algorithm of spectral projection gradient for L1 minimizition （SPGL1） was used to realize basis pursuit denoising. The result shows the sparse decomposition denoising of the actual drilling seismic reference signal is realized, the signal to noise ratio of the data is improved, and the source signal of the drill bit is recov- ered by this method.

关 键 词：勘探地球物理学 随钻地震 参考信号 稀疏理论 过完备原子库 噪声压制 

分 类 号：P631.4[天文地球—地质矿产勘探]