基于无监督学习地震信号残差恢复方法.pdfVIP

table{border-collapse:collapse;}table,th,td{border:1pxsolid#000;}

Anunsupervisedlearningmethodfor

residualseismicsignalrecovery

ChengboLi*,YuZhangandCharlesC.Mosher,ConocoPhillips

SUMMARY

Noiseattenuationhasbeenalong-standingprobleminseismicdataprocessing.Itpresentsunique

challengesonlandduetoacomplexnearsurfacecoupledwithunavoidableenvironmentalnoise

sources.Inmanycases,weaksignalisembeddedinmuchstrongernoise,whichmakes

conventionalmethodslesseffectiveonextractingthosesignals.Machinelearninghasshownits

g